Datasets:

CleverShovel

/

paper_reviews

like 1

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: FAST AND ACCURATE FACE RECOGNITION SYSTEM USING MORSCMS-LBP ON EMBEDDED CIRCUITS Review round: 2 Reviewer: 1

Basic reporting: no comment Experimental design: no comment Validity of the findings: no comment Additional comments: All of my concerns have been addressed. This version may be considered to be accepted.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: FAST AND ACCURATE FACE RECOGNITION SYSTEM USING MORSCMS-LBP ON EMBEDDED CIRCUITS Review round: 2 Reviewer: 2

Basic reporting: no comment Experimental design: no comment Validity of the findings: no comment Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: RULE-RANKING METHOD BASED ON ITEM UTILITY IN ADAPTIVE RULE MODEL Review round: 1 Reviewer: 1

Basic reporting: Author proposed novel method to find interesting patterns using lift ratio which incorporates both utility and frequency of itemsets in a database. The major highlight of this article is that authors proposed rule ranking method to extract utility based association rules. The article is structured and written well. Though this work addresses the issues of the existing work in utility based data mining, few corrections needs to be done. 1.1. Introduction should clearly addresses the issues of existing works. 1.2. The comprehensive review on utility based data mining should be done. The following recent references should be investigated and cited. i) https://www.inderscienceonline.com/doi/abs/10.1504/IJITM.2015.066056 ii) https://www.tandfonline.com/doi/full/10.1080/08839514.2014.891839 iii) http://www.cai2.sk/ojs/index.php/cai/article/view/1333 iv) https://ieeexplore.ieee.org/abstract/document/6416812/ v) https://link.springer.com/article/10.1007/s11036-019-01385-6 vi) https://link.springer.com/article/10.1007/s12652-020-02187-5 vii) https://publications.waset.org/9997900/a-distributed-approach-to-extract-high-utility-itemsets-from-xml-data viii) https://link.springer.com/article/10.1007/s11063-022-10793-x 1.3. Equations should be represented by using equation numbers. 1.4. The flow of algorithm should be clearly explained. 1.5. The algorithm needs to be written clearly with necessary indentations. Experimental design: 2.1 Experimental results needs to be explained elaborated manner. 2.2 Dataset description is not available. 2.3 Experimental results should be supported with necessary graphs by considering various factors such as number of rules generated, support, utility, confidence etc. Validity of the findings: 3.1 The validity and the findings is not clearly explained. 3.2 Statistical analysis can be performed. Additional comments: The proposed work can be explained with suitable example. Author can use small dataset for this explanation.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: RULE-RANKING METHOD BASED ON ITEM UTILITY IN ADAPTIVE RULE MODEL Review round: 1 Reviewer: 2

Basic reporting: No comment Experimental design: No comment Validity of the findings: No comment Additional comments: The authors proposed a method for ranking the rules based on the lift ratio value, which was derived using the item's frequency and utility. I would ask the authors for clarification of some issues before acceptance of this paper. Therefore, I recommend a revision. • Please, you should add a comparative study. A comparison between your algorithm and others should be added. • The limitation(s) of the association rules mining methodologies proposed in this work should be extensively discussed. • When the pseudo-codes of the proposed method are examined, it is seen that concepts such as D, S, X, and U are written in a different language. Please write them all in the same language and give pseudocodes as "appendix". • Complexity analysis should be done. • In Figure 1 "eksternal" should be fixed as "external" • You are using a specific database in Table 1. Please indicate your reasons for choosing these datasets. • Number of transactions is small in all datasets. A larger data set should be added and analysis should be performed on large data sets. • I'd like to see a more detailed analysis of the proposed algorithm's scalability. What are the main theoretical and practical benefits of the proposed algorithm? What about memory consumption and problem dimensions (especially big data)? • "II.2 and II.3" used in " Tables II.2 and II.3 show that the number… " expression should be corrected. • Is the support value given as a percentage in the test results? Not mentioned in the article? • The authors mentioned some studies in the literature (such as ELECTRE, ELECTRE II, AHP). They should also compare the proposed method with the existing methods in the literature. • The properties of the datasets used are already given in Table 1. It is given again in Table 2. If it will be given in Table 1 separately, it should be removed from Table 2. Table 1 is not needed if Table 2 will also be given.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: RULE-RANKING METHOD BASED ON ITEM UTILITY IN ADAPTIVE RULE MODEL Review round: 2 Reviewer: 1

Basic reporting: Authors made significant efforts in improving the manuscript. But, few references i have suggested which are closely associated with their work are not studied and cited. Hence, I suggest authors to study and revise the paper accordingly. Experimental design: Authors made significant efforts in improving the manuscript. But, few references i have suggested which are closely associated with their work are not studied and cited. Hence, I suggest authors to study and revise the paper accordingly. Validity of the findings: Authors made significant efforts in improving the manuscript. But, few references i have suggested which are closely associated with their work are not studied and cited. Hence, I suggest authors to study and revise the paper accordingly. Additional comments: Authors made significant efforts in improving the manuscript. But, few references i have suggested which are closely associated with their work are not studied and cited. Hence, I suggest authors to study and revise the paper accordingly.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A MULTI-STAGE HEURISTIC METHOD FOR SERVICE CACHING AND TASK OFFLOADING TO IMPROVE THE COOPERATION BETWEEN EDGE AND CLOUD COMPUTING Review round: 1 Reviewer: 1

Basic reporting: The variables need be well defined: 1. The authors defined different variables with multiple letters, i.e., in_i, map_{i,j}, bw_j, nft_i, etc, which are confusing. Commonly, ‘in_i’ in an equation means ‘i’ * ‘n_i’. Therefore, try to define a variable with a single letter. 2. There are multiple variables used in this manuscript. Adding a notation table to list all symbols/notations could improve its readability. 3. There is a highly related work [R1] on service caching and task offloading in MEC, which should be well cited. [R1] "Joint Optimization of Service Caching Placement and Computation Offloading in Mobile Edge Computing Systems," IEEE Transactions on Wireless Communications, vol. 19, no. 7, pp. 4947-4963, July 2020. Experimental design: The performance evaluation of the proposed algorithm is not sufficient. 4. With respect to the contribution of this work, since the proposed algorithm MSHCO is a heuristic algorithm, what is the performance gap between MSHCO and the global optimal algorithm? The numerical studies only show that MSHCO outperforms some existing benchmarks. However, we have no idea how good enough is the algorithm, which is essential for potential readers who want to continue studying this topic. Better provide an upper bound or the exhaustive search optimal. 5. This work considers head deadline requirements of tasks, where each task is associated with a deadline d_i. How do you handle the tasks if the deadline cannot be met? Will these unfinished tasks resources for the proposed MSHCO? How about other benchmark algorithms evaluated in the simulations? 6. In the performance evaluation, besides the comparisons with other benchmarks, the analysis on the structure of MSHCO is necessary. For example, how do these three stages of the algorithm affect its performance? Validity of the findings: no comment Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A MULTI-STAGE HEURISTIC METHOD FOR SERVICE CACHING AND TASK OFFLOADING TO IMPROVE THE COOPERATION BETWEEN EDGE AND CLOUD COMPUTING Review round: 1 Reviewer: 2

Basic reporting: Correct the punctuation error in line 37, used comma two times On line 56, grammar error “But these works have some issues must be address” The sentence on line 69, 70, 71, there is single sentence which should be avoided. In Line 87, “The section” should be replaced with “The next section” or “The second section” Line 96 have also grammar problem, should be rectified Please define the sentence on line 199, “and exams the next task” Another grammar mistake on line 199-200-201 “Otherwise, MSHCO tries to rented a new CS”. “MSHCO rents a new CS with found type” I think whole paragraph from line 196-205 should be revised because of grammar problems. Another grammar issue on line 208-209 “There are two situations when offloading a task to an ES, the requested service has or 209 not cached on the ES” Please also revise sentence on line 217-218. There are no snapshots regarding experimental work, moreover authors should compare their results in tabular format. Please also elaborate the comparison of user satisfaction in more detail. Experimental design: In this paper, the authors have proposed a method to address the joint service caching and task offloading problem in edge-cloud computing environments, to improve the cooperation between edge and cloud resources. They have proposed the method in three steps including formulating the problem into a mix-integer nonLinear programming followed by proposing a a three-stage heuristic method for solving the problem in polynomial time. Finally , they focused on improving the performance of tasks offloaded to the cloud, by re-offloading some tasks from cloud resources to edge resources. Following are my major questions It seems to be a stand alone study without the comparison with any state of the art, please try to justify your method with a concrete comparison with some similar kind of latest study. The same should also be discussed in the abstract. Please briefly justify the potential benefits of the proposed system. Please make sure to cite the references for the equations if they are not owned by you What are the formation basis and correctness justifications of Algorithm 2 & 3 ? Because the cloud & edge computing is the hot area of research these days and there are a lots of test beds available, therefore, authors are strongly recommended to perform a real study instead of the simulations of proposed model. This way , they will have a better chance to prove the authenticity of their work. Validity of the findings: It seems to be a stand alone study without the comparison with any state of the art, please try to justify your method with a concrete comparison with some similar kind of latest study. The same should also be discussed in the abstract. What are the formation basis and correctness justifications of Algorithm 2 & 3 ? Because the cloud & edge computing is the hot area of research these days and there are a lots of test beds available, therefore, authors are strongly recommended to perform a real study instead of the simulations of proposed model. This way , they will have a better chance to prove the authenticity of their work. Additional comments: please make sure to remove the language and grammatical errors through out the manuscript.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ENHANCED GENETIC FOLDING ALGORITHM FOR PROSTATE AND BREAST CANCER DETECTION Review round: 1 Reviewer: 1

Basic reporting: The manuscript has presented a clear meaning and has utilised professional English at most of the contents. However, there are concerns regarding the methodology and dataset: 1. how is the Genetic Folding algorithm developed in terms of its novelty, targeted issues in the dataset? 2. The dataset appears to be small rather than comprehensive to reflect the effectiveness of the algorithm. One another better dataset example can be: "Using deep learning to enhance cancer diagnosis and classification" 3. The figures could not exclusively identify the details for the paper. 4. The algorithms could not be translated properly with its current form. 5. Please consider to make the reference clickable. For some reasons, I could not find the reference for Table 5. Experimental design: Overall, it appears to be a work representing the quality of student project report, which has sufficiently reflected the terms and knowledge for building classifiers for particular dataset. However, it fails to reflect the studied research questions, particularly it is difficult to understand the challenge to build a qualified and SOTA classifier for the given dataset. Meanwhile, it lacks sufficient details concerning the novelty and completeness of the utilised method. It appears to be an application work. Validity of the findings: It is not strong enough to support the conclusion as SOTA performance for the given dataset. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ENHANCED GENETIC FOLDING ALGORITHM FOR PROSTATE AND BREAST CANCER DETECTION Review round: 1 Reviewer: 2

Basic reporting: In this MS the authors apply an evolutionary Genetic Folding (GF) algorithm to the binary classification of prostate cancer malignancy from clinical tumour features. The development and improvement of tumour stratification algorithms is an active area of research where advances might strongly benefit cancer diagnosis and treatment. The authors provide a clear and concise overview of prostate cancer diagnosis and grading approaches, both clinical and computational. The authors then propose and implement a SVM with GF-kernel based prostate cancer classifier (GF-SVM), arguing that GF has been shown to outperform other evolutionary algorithms. A few passages should be edited using more appropriate terminology (e.g. line 174 should read "Instances were classified as..."). There are a few typos I would recommend addressing before publication (e.g. repetition at lines 78-78, line 134 should reference Mezher et. al 2010). Experimental design: no comment Validity of the findings: The authors state (lines 205,206) that GF-SVM "demonstrated a significant performance improvement over the existing models in this domain." However, the authors have failed to perform any sort of statistical analysis to demonstrate that their method is significantly better, or even different from, the established methods (e.g. see [1]). I would recommend the authors included a pairwise analysis of all alternative models in Fig. 2C versus GF-SVM (e.g. via a binomial McNemar test [2] over at least 10 holdout shuffle replicates). Without this I do not believe this manuscript meets the "statistically sound" criterion and I cannot recommend it for publication. [1] Nicholls, Anthony. "Confidence limits, error bars and method comparison in molecular modeling. Part 1: the calculation of confidence intervals." Journal of computer-aided molecular design 28.9 (2014): 887-918. [2] McNemar, Quinn, 1947. "Note on the sampling error of the difference between correlated proportions or percentages". Psychometrika. 12 (2): 153–157 Additional comments: The authors collect and made available a tabular dataset of 8 clinical features across 100 prostate cancer patients, as well as a repository for the Genetic Folding Open source library. Unfortunately, they do not provide the reader with any way of rapidly and conveniently reproducing their models and results (e.g. via a colab or jupyter notebook). I believe this would greatly help their model's chances of being re-used by others in the community.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ENHANCED GENETIC FOLDING ALGORITHM FOR PROSTATE AND BREAST CANCER DETECTION Review round: 2 Reviewer: 1

Basic reporting: Thanks for the revision efforts. It is found that more details have been added, however, it lacks a section of describing the distinguishable technical contributions. Moreover, the reference is very limited. Please consider including more up-to-date reference, such as: Weakly supervised prostate tma classification via graph convolutional networks; Integrating genomic data and pathological images to effectively predict breast cancer clinical outcome; Supervised machine learning model for high dimensional gene data in colon cancer detection; A survey on machine learning approaches in gene expression classification in modelling computational diagnostic system for complex diseases; OncoNetExplainer: explainable predictions of cancer types based on gene expression data; A Novel Statistical Feature Selection Measure for Decision Tree Models on Microarray Cancer Detection; Selecting features subsets based on support vector machine-recursive features elimination and one dimensional-Naïve Bayes classifier using support vector machines for classification of prostate and breast cancer. Experimental design: Please refer to the mentioned reference regarding the experimental design. The p-value will be desired to evaluate the effectiveness of the model. Validity of the findings: It could be validated if the source code and datasets are publicly available. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ENHANCED GENETIC FOLDING ALGORITHM FOR PROSTATE AND BREAST CANCER DETECTION Review round: 2 Reviewer: 2

Basic reporting: The authors have expanded their model performance analysis by including a pairwise model accuracy comparison among a set of alternative SVC kernels and their proposed GF kernel in Fig. 2C,E. This analysis shows that GF outperforms all other evaluated kernels on these classification problems. I believe the manuscript now satisfies this journal's criterion for publication. However, I would personally still recommend the authors included some sort of pairwise statistical test in this analysis (e.g. Mann-Whitney, McNemar) to robustly assess the significance of the observed improvement in performance. Experimental design: no comment Validity of the findings: no comment Additional comments: no comment

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: PPE DETECTOR: A YOLO-BASED ARCHITECTURE TO DETECT PERSONAL PROTECTIVE EQUIPMENT (PPE) FOR CONSTRUCTION SITES Review round: 1 Reviewer: 1

Basic reporting: The paper is, in general, written well, used formalism is explained, and the problem is stated clearly. Also, the application seems to be interesting. Unfortunately, there are three points that need to be addressed. a) Abstract: the sentence "For the detection algorithm, this study has used the You Only Look Once (YOLO) family's anchor-free architecture, YOLOX, which yields better performance than the other object detection models within a satisfactory time interval." is not valid. The study does not bring a comparison with other SOTA detectors. The only comparison is made with YOLOv5-x. Based on general knowledge and benchmarks such as https://paperswithcode.com/sota/object-detection-on-coco benchmark, scaled YOLOv4 should yield better performance than YOLOx. The currently best detector is Swin, which still performs with a reasonable time. b) Section 4: there is no motivation to implement your own haze/rain/low-light augmentations when all of them and many more are available through Albumentations? See https://albumentations-demo.herokuapp.com/ Based on it, Section 4 is useless and can be removed. c) Section 5: it describes widely used evaluation metrics that are a gold standard in object detection and are well known by researchers. Therefore, there is no need to explain it in detail. It is enough to state you use IOU and mAP; thus, the section can be removed. Experimental design: no comment Validity of the findings: Here, it is necessary to state that two major issues were found. It is needed to address them carefully, mainly the second issue. a) Repository: - There is no readme in the repository regarding the described functionality. There is only a general readme forked from the original YOLOx repository. It must be described how to run the training script (which one is it) to replicate your results. - I did not find training data or link to them. - The model_test.ipynb functionality in the repository is nothing but a long error report. In the end, there is a test on seven images only without their visualization or mAP evaluation. In summary, the correctness of the scheme cannot be confirmed. b) Data: there is a question about the correctness of the labels. I examined several images from 'test-n.zip' and observed that: - gettyimages-88655418-612x612_jpg.rf.6f0a5bd6a8ac6cf7a77d8a24b814e8df there are two person, two glasses, two helmets, and two vests. Label includes only one glass and one person. - ppe_0579_jpg.rf.17a45729452bdb75971919d2630f0e21.jpg one orange helmet and one head are missing in the label - ppe_0994_jpg.rf.98234bc61dda7a1fb4e54e7e02ade2a0.jpg the label for the bottom-right person is missing - vitolda-klein-lAqSzwr5eQc-unsplash_jpg.rf.774701a69123aa712e43b4fc4deb1ed0.jpg the label misses orange helmet - 000221_jpg.rf.8dff87700f9373a3eff79e1f9f55f273.jpg label person is missing for all people. That holds for all (8) images from this 'subseries.' - gettyimages-83455052-612x612_jpg.rf.ea5cd5e8c3044737116b908a6139fd6b.jpg the label misses two helmets ...and many more. Based on this fact, the results published in the papers are not trustworthy because they are computed for a highly noisy and inconsistent dataset. The dataset must be fixed, and the whole experiment section must be recalculated. Additional comments: There are three minor comments: a) Is 'Md. Ferdous' full name? b) The sentence "In construction sites more than 71% injury appears than 32 in all other industries." needs to be rephrased. c) Formula 4: there is written 'Where, G and P are the prediction and ground truth bounding boxes respectively'. It is formally correct, but it is better to mark P as prediction and G as ground truth and not contrariwise.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: PPE DETECTOR: A YOLO-BASED ARCHITECTURE TO DETECT PERSONAL PROTECTIVE EQUIPMENT (PPE) FOR CONSTRUCTION SITES Review round: 1 Reviewer: 2

Basic reporting: This review paper demonstrates a YOLO-based architecture to detect personal protective equipment for construction sites. The structure of the article is organized well however, some changes need to be done to improve it. Some suggestions for further improvement: The Introduction section is well-written, and I propose to schematize all the discussion in several paragraphs (to facilitate the reading): (1) motivations, (2) the overall approach, (3) main contributions. However, the main challenges to the field and the needs and benefits of this study are missed in the introduction. A general discussion of the limitations and expectations of the proposed model should be inserted. Experimental design: Adding some effects such as haze or low light on images, will not change the features of the image significantly. I suggest creating a night-time dataset and adding it to the existing one. The authors mentioned that the effected images are used only for testing. What about the original images of the effected images? If the same original images are used for training, then testing with the same images with just a small effect is not a good idea because in that case, the model will just memorize the images. Please specify what is the main differences between CHV and CHVG datasets. Is it the same datasets just the number of classes are increased or the HAZE, RAIN, and LOW-LIGHT images are included as well? Has exactly the same YOLOX model been used in this article or have the authors made any changes to it? This needs to be explained in the paper clearly. Line 120: there are two “also” which are not necessary and can be removed. Line 146: Please explain why DarkNet53 has been used as pre-trained weights in this model. Line 186 and 188: Authors need to determine why they set these parameters. Line 194: The image size needed to be explained in the dataset section. Line 259: Please determine why the NMS value is set to 0.65 Validity of the findings: The evaluation metrics section is explained very well, however, the discussion section is too short and needs to be expanded. It would be good if authors upload their code on GitHub and add the link in the paper. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: PPE DETECTOR: A YOLO-BASED ARCHITECTURE TO DETECT PERSONAL PROTECTIVE EQUIPMENT (PPE) FOR CONSTRUCTION SITES Review round: 2 Reviewer: 1

Basic reporting: I appreciate that the authors carefully processed all my comments and improved the paper significantly. It is clear, easily readable, and correct. Experimental design: I have no critical comments in this section. Validity of the findings: The authors fixed the issue with the dataset labels. The findings are now supported by the data and I have no more critical comments. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: PPE DETECTOR: A YOLO-BASED ARCHITECTURE TO DETECT PERSONAL PROTECTIVE EQUIPMENT (PPE) FOR CONSTRUCTION SITES Review round: 2 Reviewer: 2

Basic reporting: No Comment Experimental design: No Comment Validity of the findings: No Comment Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: THE ROLE OF METACOGNITION IN PROMOTING DEEP LEARNING IN MOOCS DURING COVID-19 PANDEMIC Review round: 1 Reviewer: 1

Basic reporting: The paper aims at addressing role of metacognition in promoting deep learning via MOOCs. The work targets to answer two research questions. The work reported though not novel is acceptable contribution as the attempt is considerable. Experimental design: Three aspects Critical thinking, Connecting concepts and Creating new concepts are considered as evaluation strategy. The questionnaire enclosed along with manuscript is considered for review. In my view, A detailed analysis w.r.t Anova would have been better instead of just stating the outcome. Some more w.r.t analysis and key findings would add value. Authors can address these in revision. Validity of the findings: Computation enclosed in excel file support for validation. Conclusion is relatable to discussion done in the paper. The approach though not novel is acceptable attempt. Additional comments: A more comparative evaluation is preferred as present version is limited. Usually, analysis is expected but analysis part is very limited These two points may be considered

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: THE ROLE OF METACOGNITION IN PROMOTING DEEP LEARNING IN MOOCS DURING COVID-19 PANDEMIC Review round: 1 Reviewer: 2

Basic reporting: its an attempt to explore the Role of Metacognition in Promoting Deep Learning in MOOCs during COVID-19 Pandemic. deep learning is not elevated in detail. data set taken for experimentation is not mentioned in detail. Experimental design: needs more explanation about dataset taken and its features. how they have trained the model and how they tested. utilization of deep learning model to be discussed. Results generated from the proposed idea are limited Validity of the findings: somewhat okay but dataset size and features to be enhanced. model should be more scalable with more variety of data and with higher dimensions are able to process in dynamic platform. Additional comments: more investigation to be done on idea(methodology), dataset taken and projecting the role of deep learning in MOOCs

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: THE ROLE OF METACOGNITION IN PROMOTING DEEP LEARNING IN MOOCS DURING COVID-19 PANDEMIC Review round: 1 Reviewer: 3

Basic reporting: As per my observation, the studies were done on covid -19 and are related to the title given. Participants were students at the department of home economics who were all at the seventh academic level. Based on their scores on the metacognition awareness inventory, they were divided into two experimental groups, i.e high metacognition students and low metacognition students done with a small sample. if it is taken for a large sample will the same results will be appeared? Experimental design: As per the Author's view 203 out of 260 quantitative MAI points, metacognition was graded into two groups, high 204 metacognition (HM ≥ 65 percent) and low metacognition (LM < 65 percent) in line with Aydın205 and Coşkun (2011); Redondo and López (2018). In other words, based on their scores on the 206 MAI instrument, participants were divided into two experimental groups. There is a scope for comparative study. is it possible to incorporate? Validity of the findings: is it justifiable that the authors noted that all participants have studied and passed the course's prerequisite 14 courses, mainly Research Methodology, Teaching & Statistics Principles, and Principles of 215 statistics. Performance can be upgraded Additional comments: Check the Grammar again

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: THE ROLE OF METACOGNITION IN PROMOTING DEEP LEARNING IN MOOCS DURING COVID-19 PANDEMIC Review round: 2 Reviewer: 1

Basic reporting: All responses are addressed in the revised version of the manuscript. Experimental design: Investigation is carried and method described are sufficient w.r.t revised version of manuscript. Validity of the findings: Findings are validated with proper data. Additional comments: All review comments are answered and incorporated appropriately.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: THE ROLE OF METACOGNITION IN PROMOTING DEEP LEARNING IN MOOCS DURING COVID-19 PANDEMIC Review round: 2 Reviewer: 2

Basic reporting: manuscript is well organized and explored on important point required in current context Experimental design: methodology given well and implemented too to obtain specific outcomes Validity of the findings: findings obtained from the proposed method are valid may be improved in the performance on various datasets/various data sizes Additional comments: manuscript will be so worthy if mentioned Limitations are reduced and future directions are implemented further

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: THE ROLE OF METACOGNITION IN PROMOTING DEEP LEARNING IN MOOCS DURING COVID-19 PANDEMIC Review round: 2 Reviewer: 3

Basic reporting: Yes, the revision version is fine. Experimental design: The results enclosed in revision are fine Validity of the findings: Supporting data is provided for validation. Additional comments: The revision version is acceptable.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: REAL-TIME DDOS FLOOD ATTACK MONITORING AND DETECTION (RT-AMD) MODEL FOR CLOUD COMPUTING Review round: 1 Reviewer: 1

Basic reporting: The analysis of the findings is very interesting. Many challenges of this proposed method are covered. Literature reference is relevant, but does not cover all the topics, such as AI, and is not so up to date. Experimental design: As a concept, it is well designed. I found the topic interesting and definitely an area of growth. Validity of the findings: Despite the research effort and the achieved high rate of accuracy, there are a number of challenges that need to be addressed, such as execution-time for online learning of random forest algorithm, the lack of up-to-date real-world datasets for training, what is the rate of false-positive and false-negative, there are scalability issues or performance? Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: REAL-TIME DDOS FLOOD ATTACK MONITORING AND DETECTION (RT-AMD) MODEL FOR CLOUD COMPUTING Review round: 1 Reviewer: 2

Basic reporting: The paper is well structured and easy to read. Also, the use of English is quite good. The introduction provides a great, generalized background of the topic and the motivations for this study are clear. A minor comment concerns the References part, where it would have been better to have used more recent literature with more up-to-date knowledge. The experimental part is well written and is appropriate for the study. To conclude, this research work apparently fulfills the purpose for which it was carried out. It would be interesting to see in the future, the use of more machine learning classifiers in building the Real-Time DDoS flood Attack Monitoring and Detection RT-AMD predicting model or even use more datasets for executing the same experiments. For the above reasons I strongly recommend the acceptance of this paper. Experimental design: The research question of this research is well defined and undoubtedly adds knowledge to an otherwise quite explored scientific area, that of attack monitoring and detection in cloud computing. The methods used, although not innovative, have certainly been used in the right way and the overall investigation is performed to a high technical standard. Validity of the findings: Throughout the research there is no replication of pre-existing knowledge. Conclusions are both well stated and linked to original research question. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: SOLVING THE CLUSTERED TRAVELING SALESMAN PROBLEM VIA TRAVELING SALESMAN PROBLEM METHODS Review round: 1 Reviewer: 1

Basic reporting: In this paper, the authors investigate the clustered traveling salesman problem (CTSP), which is an extension of the popular TSP. The paper is clear and the conclusions are supported by extensive computational results. I do recommend to add at the references the following paper: P.C. Pop, I. Kara and A. Horvat Marc, New Mathematical Models of the Generalized Vehicle Routing Problem and Extensions, Applied Mathematical Modelling, Elsevier, Vol. 36, Issue 1, pp. 97-107, 2012. which presents some interesting MIP formulations of the CTSP, and the papers: O. Cosma, P.C. Pop and L. Cosma, An effective hybrid genetic algorithm for solving the generalized traveling salesman problem, Lecture Notes in Computer Science, Vol. 12886, pp. 113-123, 2021. P.C. Pop, O. Matei and C.M. Pintea, A two-level diploid genetic based algorithm for solving the family traveling salesman problem, in Proc. of GECCO 2018, Association for Computing Machinery, Kyoto, Japan, 2018. These papers deal with problems closely related to CTSP: the generalized traveling salesman problem (GTSP) and the family traveling salesman problem (FTSP). It is important to mention these papers in the introduction, especially because the authors are using GTSP instances. Experimental design: The authors used the transformation suggested by Chisman in order to transform the investigated problem into classical TSP, which then is solved using state-of-the-art algorithms. The research questions are well defined and they are relevant. Validity of the findings: The conclusions are well stated and they are clearly linked to the proposed research questions. Therefore, taking into consideration all the mentioned aspects I do recommend to accept the paper for publication if the authors answer to the addressed comments. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: SOLVING THE CLUSTERED TRAVELING SALESMAN PROBLEM VIA TRAVELING SALESMAN PROBLEM METHODS Review round: 1 Reviewer: 2

Basic reporting: This manuscript investigates methods for solving the Clustered Traveling Salesman Problem (CTSP), specifically, by converting it to the Traveling Salesman Problem (TSP) that has been extensively studied. The experimental results essentially answer several of the questions proposed by the authors. Experimental design: no comment Validity of the findings: no comment Additional comments: There are several issues below that need further clarification by the authors: 1. In Section 1, in addition to introducing the literature involved in the relevant methods, the characteristics and scope of application of the corresponding methods should also be summarized. 2. Again in the Section 1, some references seem to be too old. 3. The focus of this paper is to explore the conversion of CTSP to TSP for solving. The conversion method adopted by the author and the considerations behind it should be described in more detail in Section 2.1. 4. On page 8, line 285, what does “Err” in “B-Err” stand for? Authors may consider using more appropriate abbreviations. 5. The author's research found that several TSP solvers have different performance for CTSP. However, it would also make sense to discuss the reasons behind these differences in further depth. 6. When comparing the results, some statistical methods can be added to compare the superiority of several methods, such as variance analysis, Friedman test, etc.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: SOLVING THE CLUSTERED TRAVELING SALESMAN PROBLEM VIA TRAVELING SALESMAN PROBLEM METHODS Review round: 2 Reviewer: 1

Basic reporting: The authors took into consideration all the comments and observations and as a consequence, I do recommend to accept the paper for publication. Experimental design: No comment. Validity of the findings: No comment. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: SOLVING THE CLUSTERED TRAVELING SALESMAN PROBLEM VIA TRAVELING SALESMAN PROBLEM METHODS Review round: 2 Reviewer: 2

Basic reporting: most of the comments have been addressed, just need to do proofreading before publication Experimental design: most of the comments have been addressed, just need to do proofreading before publication Validity of the findings: most of the comments have been addressed, just need to do proofreading before publication Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: KINSHIP IDENTIFICATION USING AGE TRANSFORMATION AND SIAMESE NETWORK Review round: 1 Reviewer: 1

Basic reporting: - Problem statement, solution provided and outcome should be clearly explained in abstract. The flow of information should be in order and connected. - Description of big data in the beginning of introduction is not relevant and should be removed. Start from pictorial discussion. - Clearly explain the problem statement in the introduction section. - Contribution of paper mainly highlights the methodology not the contribution. Explain how the results and outcome of this research contributes to the body of knowledge. - Remove paper organization section. - "Joseph P.R. et al Introduced....." use standard style of reference citation in the text. - "resemblance [8]. Zhang et al. (2015)" - follow only one style of referencing throughout the paper. - Literature should be written in discussion form not simple descriptive form. - "we present a model....." don't use I, we, our in research. use passive voice sentences. Experimental design: In proposed work, use modeling form (graphical/mathematical etc) to explain the methodology and explain all the steps in sequence. Give research model. - Support the methodology with references. - Provide details of data set, its source and how much data is used for training and implementation? - Equations are inline with text. Should be on separate lines, with standard equation format and numbering, and explain them. - Properly define the design of methodology and explain it step by step. Validity of the findings: - How the validity of the results checked? - Explain results and show them with the proper technical format rather than simply mentioning them. - Use tables, data, numbers and figures to explain the results. - Data given in tables should be well explained and justified. - How results are significant? - Conclusion should explain how results support the objectives. what is the significance of the study. Summarize the outcome of the research, its impact and applicability. Additional comments: - Make grammatical corrections.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: KINSHIP IDENTIFICATION USING AGE TRANSFORMATION AND SIAMESE NETWORK Review round: 1 Reviewer: 2

Basic reporting: no comment Experimental design: a)- What was the specific reason for picking images of an age between 15-20 only? b)- What was the dataset size used in the experimental validation. c)- Why CNN-based deep relational network was used for features extraction from the facial images of the dataset? Validity of the findings: a)- Why the proposed model is taking 10 images of each member? why not 5 or 15 or more? b)- Any specific reason why Father-Daughter and Mother-Son accuracy in both Training and Validation Accuracy is lower as compared to Father-Son and Mother-Daughter? c)- Have we also compared any other performance factor other than accuracy from the other proposed models mentioned in Table1. Additional comments: We can only say that the proposed model's performance is better than the previously proposed model (Table1). We cannot say that the previously proposed models failed to deliver improved performance.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: KINSHIP IDENTIFICATION USING AGE TRANSFORMATION AND SIAMESE NETWORK Review round: 1 Reviewer: 3

Basic reporting: English language corrections are needed at some places in the article. Proofreading is recommended. E.g. line 194: Spelling mistake ‘pre-processing’, Line 284: use figure 4 instead of diagram 4 Experimental design: no comments Validity of the findings: no comment Additional comments: Repetition of the basic working is observed in multiple sections. It is suggested that more details be added instead of repetitions of the basic working. Theme of Figure 3 and figure 4 is the same. Care must be taken when using abbreviations. It is suggested that the full form and its abbreviation be provided at the first appearance. Afterward only abbreviated form be used. E.g. RFIW –Full form and its abbreviation should be provided at first appearance only. Afterward only abbreviated form should be used. Lines 62, 82, 315. Similarly, at line 324 only the abbreviation LAT should have been used. Many other such examples are present. All equations must be numbered and cited in the text.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM READOUT AND GRADIENT DEEP LEARNING MODEL FOR SECURE AND SUSTAINABLE DATA ACCESS IN IWSN Review round: 1 Reviewer: 1

Basic reporting: 1.Use of deep learning models with wsn is good for intrusion detection 2. For authentication and authorisation alone ML/DL are not required, however to minimise the false acceptance rate use of sophisticated algorithms is good technique Experimental design: 3. It is generally preferable to use light weight techniques in wsn/ IoT systems , deep learning is time and resource intensive when it comes to training. If a certain protocol has to be followed for authentication ,violation of it leads to an alarm. To detect this violation an anomaly based model is used . The same can be implemented without a model by simply using a checklist that contains a exhaustive list of items that need to be covered before the device can be safely added to the network. So while the approach of the author seems to be effective for now , better algorithms will keep coming up. To setup an authentication based on checklist mechanism is however time resilient unlike a learning based method. Statistical analysis of data could be useful from time to time , but for wsn lesser data that needs to be processed the better. 4. Deep learning and machine learning models tend to work based on historic data and behaviour . Training the models when planned changes occur in the infrastructure is an operational overhead. Validity of the findings: 5 gradient based algorithm is not the best approach in security mechanisms as , things need to be as deterministic as possible . Sometimes the position of a sensor or its configuration could be constantly changing in the IWSN . In such a case it's not clear how this system works. Additional comments: No comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM READOUT AND GRADIENT DEEP LEARNING MODEL FOR SECURE AND SUSTAINABLE DATA ACCESS IN IWSN Review round: 1 Reviewer: 2

Basic reporting: The English language of the paper is clear and readable. • There is a need to explain the authentication problem in more details, some of the IIoT networks are designed in a closed network where they have their own cables and devices and no one outside the network has any access to it, in this case the authentication problem needs to be clarified at what circumstances it is considered an issue. I suggest the authors provide an example to explain when/how the authentication step should be included in the design. • Line 67: “in this paper” is repeated. • Authors need to identify the terminologies earlier in the paper such as authentication rate, authentication acceptance false rate, and authentication time. • Figure 1 in the paper should me improved to tell more details about the system model; not much information can be extracted from the figure in its current design. • Line 213: “the gateway node has to approve the authenticity of the sensors. The gateway node verifies whether the sensor has been tampered with or not in order to guarantee the confidentiality and integrity of the data packets”. Authors are required to identify what the gateway node is, what type of data is gathered, and how the authentication, confidentiality, and integrity of the sensors are being verified? • Line 220: Extra word “input” • Figure 2 is meaningless and not clear, E.g., What is the input look like? Etc.,. • The paper has no example of where their design is applicable. • Line 237, “The QSAE checks for energy availability, authenticity and authorization of the sensor by employing the quantum readout function to ensure smooth communication between the sensor and the supervisory control unit with a minimum false acceptance rate”. How are the energy availability, authenticity, and authorization verified by the QSAE? Authors should mention that it will be explained in coming sections. • QRH should be used in full words for the first time then used as a short cut. The approach of this paper in not fully explained with the technical and theoretical details, more details need to be explained. Experimental design: • Equation #7 should be Sj0 instead of Si0. • Line 296: The gradient-associating sensors is not identified anywhere earlier in the paper. • Equations should be after the parameters are identified for better understanding. • Why the number of the simulation runs is 10? No explanation of why this value is chosen, in the literature, values converged after 45 runs. Validity of the findings: • In all tables, the increase in values gradually with the increase in the number of sensors ; this is intuitively expected and should not be mentioned as a result. However results of optimizations are valid. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM READOUT AND GRADIENT DEEP LEARNING MODEL FOR SECURE AND SUSTAINABLE DATA ACCESS IN IWSN Review round: 2 Reviewer: 1

Basic reporting: All comment seems to be resolved except the following: • There is a need to explain the authentication problem in more details, some of the IIoT networks are designed in a closed network where they have their own cables and devices and no one outside the network has any access to it, in this case the authentication problem needs to be clarified at what circumstances it is considered an issue. I suggest the authors provide an example to explain when/how the authentication step should be included in the design. • Authors need to identify the terminologies earlier in the paper such as authentication rate, authentication acceptance false rate, and authentication time. I think authors should submit a corrected copy that marked all text change in red, it will be easier to see the new updates. Experimental design: 343-347 references and citations are required. Validity of the findings: Conclusion is not sufficient, more details should be illustrated in this section. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM READOUT AND GRADIENT DEEP LEARNING MODEL FOR SECURE AND SUSTAINABLE DATA ACCESS IN IWSN Review round: 3 Reviewer: 1

Basic reporting: All comments are resolved. Experimental design: All comments are resolved. Validity of the findings: All comments are resolved. Additional comments: All comments are resolved.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: BILINEAR POOLING IN VIDEO-QA: EMPIRICAL CHALLENGES AND MOTIVATIONAL DRIFT FROM NEUROLOGICAL PARALLELS Review round: 1 Reviewer: 1

Basic reporting: 1. The expression of the article is clear and coherent. However, there are some typos. In line 106, 'abstracted' to 'abstract'. In line 423, 'eachother' to 'each other'. Some terms are not easy to understand, such as in line 455, 'ImageNet-style feature vectors'. Many experiment results in the tables are not rigorous, i.e., the baseline offsets in Tables 3 and 5 have no percentage sign. 2. The overall article is well organized. However, the structure in the discussion section is chaotic. The analysis of experimental results may be more reasonable in the experiment section, for readers to contact contextual content more easily. Experimental design: 1. The experimental results are relatively single, only quantitative analysis results are shown. There are no more qualitative or analytical experimental results. Validity of the findings: 1. All experimental results showed no improvement from baselines, which was a little bit strange. I doubt the correctness of the experimental setup, for the BLP has its rationality as the author proposed and has obvious performance improvement in image QA tasks[1]. [1] Ben-younes, H., Cade`ne, R., Cord, M., and Thome, N. (2017). Mutan: Multimodal tucker fusion for visual question answering. 2017 IEEE International Conference on Computer Vision (ICCV), pages 655 2631–2639. 2. The insights given in the article are relatively basic, and the discussion is not deep enough. The author did not give an overall summary or analysis of the multiple BLP methods. The author gives a brief introduction of the neurological parallels, however, their theoretical inspiration is not clearly explained. In addition, the reasons why the BLP does harm on the video QA task didn't convince me very well. If it does do harm, what kind of interaction method should we use in the future for cross-modality representation need to be discussed further. Additional comments: no comment

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: BILINEAR POOLING IN VIDEO-QA: EMPIRICAL CHALLENGES AND MOTIVATIONAL DRIFT FROM NEUROLOGICAL PARALLELS Review round: 1 Reviewer: 2

Basic reporting: The paper focuses on the bilinear pooling problem in the context of Video-QA. The problem is indeed topical and subject to many research works. The topic is therefore appropriate and also important. The quality of the English language is not a problem and the paper is clearly written from both a literary and scientific point of view. The structure of the manuscript is quite in accordance with the standard required by the journal and the figures are of good quality with good definition and do not suffer from any particular problem. The datasets used are among the most appropriate and the availability of two of them is well indicated in the paper (TVQA and HME-VideoQA) but not for the others. There is no indication on the availability of the dual stream model proposed by the authors. The title of the paper however raises an interesting and important question but the content of the paper does not provide any explanation or clear answer to the question induced by the title. This is one of the weaknesses of this paper that should be improved in order to have a real added value and a significant contribution of the work. Experimental design: The subject matter is well within the scope of the paper and raises an important issue that is expressed in the title, namely "The limitations of bilinear pooling in video-QA". These limitations are indeed shown by means of experiments and comparisons with state-of-the-art methods. Unfortunately, there is no clear explanation of the underlying causes of this weakness of bilinear pooling except for a problem related to datasets, knowing that some of these datasets, such as MovieQA and YouTube2TextQA, seem, according to the literature, to provide the necessary for this kind of studies. In my humble opinion, it would be necessary to explain the current failures of bilinear pooling which may be due to a problem of alignment of textual and video stream modalities with also temporal alignment or to the expression of queries. A conclusion on this subject would also be welcome and there are probably two among which one should decide. The first conclusion is that the bilinear pooling is not adapted to the problem by explaining why, but this conclusion, in my opinion is not the right one because it is enough to have the adequate datasets to solve part of the problem. The second conclusion should propose corrections to the existing models. The ideas proposed in the paragraph "PROPOSED AREAS OF RESEARCH" (line 607) remain mostly general and vague intuitions. More concrete and forceful proposals would give more weight to the paper. The idea of demonstrating the weaknesses of a model is a very good approach in research, but again, these weaknesses should be clearly expressed, explained and alternatives or corrections proposed. Validity of the findings: The scientific findings are at this stage quite light and need to be improved. The preceding comments may guide the authors in improving their contributions. The authors claim that, to their knowledge, they are the first to describe how important neurological theories, namely the dual coding theory and the two-stream vision model, are related to the history (and background) of modern multimodal deep learning practices. The thinking is certainly good but needs more formal framework and mathematical expression to be effectively exploited. My suggestions to authors 1. Try to explain clearly where the current weaknesses of bilinear pooling lie and not just the dataset problem. 2. In the light of these weaknesses propose ideas for improvement or correction. 3. Formalize their idea of the dual coding theory and the dual stream vision model so that it can have a significant contribution in this area. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: DETECTION OF HIERARCHICAL CROWD ACTIVITY STRUCTURES IN GEOGRAPHIC POINT DATA Review round: 1 Reviewer: 1

Basic reporting: This paper presents a synthetic data generator that reproduces structures commonly found on geographical event data sets and describes clustering and hierarchical scale extraction algorithms to be evaluated. Basically, the topic is interesting and the manuscript is well-written but I have some comments to be improved. Major comments: - The authors mentioned that the second contribution is to propose an improvement of the DBSCAN algorithm, but existing clustering algorithms were introduced and used. Please clarify improvement points of the algorithm. The title should also be modified to fit the content because the current one seems to be the proposal of an algorithm. Minor comments: - Figure 1 should be explained more detail such as the meaning of colors. Also, the quality of the Figure 1(b) should be improved. - The authors discussed the results in Section 7, so the section name should be Results and discussions. - There are some typos as follows. Line 20: “even data” would be “event data” Line 166: “as shown in 1” would be “as shown in Figure 1” Line 431: There is a space missing after the comma. Line 492: “This results” should be “These results” Experimental design: no comment Validity of the findings: no comment Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: DETECTION OF HIERARCHICAL CROWD ACTIVITY STRUCTURES IN GEOGRAPHIC POINT DATA Review round: 1 Reviewer: 2

Basic reporting: 1. There are several places (Lines 131-137, Lines 186-191, Lines 494-496) where it is argued that the expected differences between different clustering algorithms (e.g., HDBSCAN and OPTICS vs. DBSCAN and its adaptive version). It would be beneficial to use a toy example dataset to demonstrate the differences argued here. 2. Some figures (e.g., Figure 1) need to be of higher resolution to be legible (check all other figures for this issue). In Figure 1 caption, please describe how (a) corresponds with (b). For example, is a node in (b) a point in (a)? 3. Lines 160-161: The reviewer did not quite understand what “three iterations of head-tail breaks” entails. Please elaborate. 4. Line 249: What important factors need to be considered when manually setting parameters? 5. Line 518: Does this "new method" refer to applying clustering algorithms recursively to discover structure within structure in general, or specifically to applying the adaptive DBSCAN recursively? Need to be clear. Experimental design: 1. Line 404: “statistically significant evaluations”. What does this entail exactly? Later in the results section, only box plots are used to show the differing performance of the algorithms. I would expect some sorts of statistical significance test on the differences (or comparison of confidence intervals?). 2. Lines 530-531: The reviewer thinks it is necessary to do so in this paper. Apply the adaptive DBSCAN a real-world dataset recursively to discover structures within structures, and see if the results make sense (probably with only qualitative evaluations). Validity of the findings: 1. Line 404: “statistically significant evaluations”. What does this entail exactly? Later in the results section, only box plots are used to show the differing performance of the algorithms. I would expect some sorts of statistical significance test on the differences (or comparison of confidence intervals?). 2. Lines 530-531: The reviewer thinks it is necessary to do so in this paper. Apply the adaptive DBSCAN a real-world dataset recursively to discover structures within structures, and see if the results make sense (probably with only qualitative evaluations). Additional comments: See the annotated pdf file attached for other minor comments.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: DETECTION OF HIERARCHICAL CROWD ACTIVITY STRUCTURES IN GEOGRAPHIC POINT DATA Review round: 2 Reviewer: 1

Basic reporting: no comment Experimental design: no comment Validity of the findings: no comment Additional comments: Thank you for your revisions. I checked the revised manuscript and confirmed that the authors have addressed my major and minor comments. I think the paper is accepted for publication after minor revision below. 1. Figure 1: Instead of arranging figures vertically, how about arranging them horizontally? 2. The caption of Figure 14: hours should be removed.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: DETECTION OF HIERARCHICAL CROWD ACTIVITY STRUCTURES IN GEOGRAPHIC POINT DATA Review round: 2 Reviewer: 2

Basic reporting: The revised manuscript has sufficiently addressed my comments. I now recommend acceptance for publication. Experimental design: The revised manuscript has sufficiently addressed my comments. I now recommend acceptance for publication. Validity of the findings: The revised manuscript has sufficiently addressed my comments. I now recommend acceptance for publication. Additional comments: The revised manuscript has sufficiently addressed my comments. I now recommend acceptance for publication.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 1 Reviewer: 1

Basic reporting: Line 87: The "Q" should be italicized. Line 147: "definition similarity": Did you mean "definition of similarity"? Line 216: Suggest to remove the word "real". Experimental design: Line 297: Is the alpha optimized for the networks in this research work, or it is an optimized value that can be used in any other network? Kindly clarify. Validity of the findings: Table 5 and 6: Is there a reason why the LPA-MNI is included in Table 6 but not in Table 5? As it is also one of the latest LPA variants. Line 380: In most cases, LPA and its variants will start to deteriorate past a certain threshold of mixing parameter. I wonder how well the proposed method handles higher values of mixing parameter. For example, mu = 0.6 and 0.7. The proposed method produces high-quality community division in terms of NMI and Q for the LFR networks. It outperformed the other methods in terms of Q (more than double in some cases). Thus, it will be interesting to see what is the number of communities that are detected by the proposed method as compared to the actual number of communities in the LFR networks. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 1 Reviewer: 2

Basic reporting: This paper proposes a new community detection method based on LeaderRank and LPA, the problem has significant importance, and the proposed method seems somewhat interesting. The related work section should include more related and recent studies to state the new contribution of the proposed method against ones. I could find a few very recent studies on the use of LPA in community detection which have not been indicated; moreover, the new contributions and shortcomings of existing methods may also be included somewhere in the manuscript? A working example to illustrate the functioning of the proposed method could be of immense help to the readers. Experimental design: The authors may consider incorporating the convergence, sensitivity to the starting conditions, and other relevant characteristics of the proposed algorithm (including time complexity analysis). The authors may compare their approach with the current benchmarks on community detection using relatively real-world data sets of significant size based on the actual CPU time (current datasets considered are too small). Validity of the findings: The proposed method should be competitive in computational performance for large-scale network data sets (>30000 nodes, for example); otherwise, the utility of the proposed approach to deal with realistically large networks shall remain unknown. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 1 Reviewer: 3

Basic reporting: no comment Experimental design: no comment Validity of the findings: no comment Additional comments: In this manuscript entitled “an improved two stage label propagation algorithm based on LeaderRank”, authors propose an improved two-stage label propagation algorithm to solve the problems of poor stability and low modularity. In the introduction, authors sum up their contributions and experimental findings. They focus their study on label propagation algorithm for community detection, they point out its advantages as low time complexity and suitability for large scale-networks as well as its limitations as randomness and instability. Authors give then a short overview of improved algorithms that take into consideration the importance of each node in order to seed node selection. They detailed the LPA-TS algorithm and underlined the inaccurate proposed similarity measurement between nodes that aims to rank nodes for labeling. They propose a new version of LPA-TS named LPA-ITSLR where an improvement of the similarity measurement is proposed as well as an optimised strategy for node label updating based on LeaderRank value, this will avoid the randomness of ordered labels update. The paper is well organised in two main other parts beside the introduction and the conclusion. In the second part, theoretical basis with the necessary notions and indicators are exposed and defined. Likewise, the proposed algorithm is detailed. The third part detailed experiment and analysis, experiments are carried out on six classical social network data sets where users show that LPA_ITSLR performed well for all these data sets and yielded more stable community partitioning in comparison with LPA and LPA-TS. Performance comparison of LPA_ITSLR with four other recent community detection algorithms are also carried out on the same classical data sets. Two of these algorithms are based on the importance of nodes. Experimental results show that LPA_ITSLR yielded the highest modularity and the most stable results. Other comparison results with seven other classical community detection algorithms show also good performance especially in term of modularity on the three following datasets: Karate, Dolphin and football. Other experiment studies were carried out on artificial sets and showed the same improvement. Here are some suggestion and remarks: The complexity of the algorithm LPA-ITSLR is not studied neither experimented on large-scale graph. In fact, considering the new similarity measurement as well as the new proposed strategy for updating labels leads certainly to improve the randomness and the instability of the algorithm, but may, on the other hand, slow down the processing in comparison to the classical LPA. Could you please comment this or mention it is as a perspective of your study? On the other hand, here are some remarks related to terminology, authors use the terms “community discovery” instead of the most common used one which is “community detection”. Is there any reason to this? In table 3, it seems to me that d is the diameter and not the path average length. Likewise, c is called “clustering coefficient” and not the aggregating one. At line 49 the place of the citation number 1 is inappropriate, could you please put it at the end of the sentence? At line 241, please check the sentence: “the node .. and the community .. has a large similarity”

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 2 Reviewer: 1

Basic reporting: No comment. Experimental design: No comment. Validity of the findings: No comment. Additional comments: The paper is well written and I am looking forward to any related future work on this method.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 2 Reviewer: 2

Basic reporting: Supplementary material contains .class files (unreadable) of most of the code/ material, which is not following the Peerj CS Policy where code should be made available ( Though it may be with restricted access). Experimental design: The current utility of the proposed method to deal with realistically large networks remains unknown. Validity of the findings: I appreciate the efforts made by the authors to address my comments in the previous round (although not all of them have been answered). Different parts of the paper have been improved. However, I still have a major concern that results from previous remarks that have not been addressed. Specifically, my comment on the "Validity of Findings" has not been satisfactorily answered. I reiterate my comment below. "...Validity of the findings The proposed method should be competitive in computational performance for large-scale network data sets (>30000 nodes, for example); otherwise, the utility of the proposed approach to deal with realistically large networks shall remain unknown....". Based on the availability of large-scale network data set, one would expect a new community detection algorithm to be competitive in terms of computational performance. The authors should explain how the proposed method is competitive for realistically large networks? Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 3 Reviewer: 1

Basic reporting: Different parts of the paper have been significantly improved in the previous rounds of review. Experimental design: I appreciate the author's effort to address my comments in the previous round, although not all of them have been answered. Additional experiments may be required to address my concerns which have not been addressed adequately. Validity of the findings: I still have a major concern that has not been addressed. It is important to show the computational competitiveness of the proposed method for realistically large networks and not for LFRs (which are used by the authors to address my comments). Can the proposed method outperform any one of the benchmarks like the Fast greedy algorithm, Walktrap algorithm, or Edge Betweenness algorithm on realistically large networks? It is not easy to understand the novelty of publication without having a clear and convincing contribution, as suggested above. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 3 Reviewer: 2

Basic reporting: no comment Experimental design: no comment Validity of the findings: no comment Additional comments: In this manuscript entitled “an improved two stage label propagation algorithm based on LeaderRank”, authors propose an improved two-stage label propagation algorithm to solve the problems of poor stability and low modularity. According to first round review suggestions, authors improved the algorithm and carry out experiments on large-scale networks. The largest data set in the original paper contains 10000 nodes. they generated simulation data sets containing 20000, 30000, 40000 and 50000 nodes respectively. From the experimental results, they draw the conclusion that, as the size of the data set increases, the time complexity of this algorithm increases, but it can still achieve a high modularity and NMI in terms of large-scale data sets. It would be preferable to give some measurement allowing to observe how the time complexity is increasing.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN IMPROVED TWO-STAGE LABEL PROPAGATION ALGORITHM BASED ON LEADERRANK Review round: 4 Reviewer: 1

Basic reporting: Seems ok Experimental design: Seems ok Validity of the findings: The authors have made efforts to address some of the major concerns raised in the previous rounds of review; I recommend this revised manuscript for further processing. Additional comments: Authors should check leftover grammatical/typographical issues (like in Table 12) at their end.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MALWARE DETECTION FRAMEWORK BASED ON GRAPH VARIATIONAL AUTOENCODER EXTRACTED EMBEDDINGS FROM API-CALL GRAPHS Review round: 1 Reviewer: 1

Basic reporting: Thias work presents and interesting research on applying GVAE to dimensionality reduction, followed by some classic ML models applied to the resulting features. Some comments: - Why to choose linear SVM and LightGBM and not others. - Highlight all assumptions and limitations of your work. - Conclusions should provide some lessons learnt. - Related works section does not mention recent research effors in feature characterization in NIDS using also clustering and contrastive learning, as well as, review works on VAEs. Authors are advised to refer to the following related articles to add some discussions: [1] Variational data generative model for intrusion detection, Knowledge and Information Systems, 2018 [2] Analysis of Autoencoders for Network Intrusion Detection, Sensors, 2021. [3] Network Intrusion Detection Based on Supervised Adversarial Variational Auto-Encoder With Regularization, IEEE Access, 2020 [4] Supervised contrastive learning over prototype-label embeddings for network intrusion detection, Information Fusion, 2022 Experimental design: The experimental design and the descriptions of experimental results is well done and sufficient. Validity of the findings: Accuracy anf F1 has been obtained, and to assess the significance of results a Wilcoxon Signed Rank Test was applied using a 0.05 significance level. The results are correctly analyzed and are significant. To obtain some additional metrics eg recall, precision, etc,...would be fine but it is not strictly necessary. Table 9 of classification results of recent malware studies is interesting to position the results in perspective with others. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MALWARE DETECTION FRAMEWORK BASED ON GRAPH VARIATIONAL AUTOENCODER EXTRACTED EMBEDDINGS FROM API-CALL GRAPHS Review round: 1 Reviewer: 2

Basic reporting: This study proposes a malware detection model based on API-call graphs and uses Graph Variational Autoencoder (GVAE), which is a variant of graph neural networks, to reduce the size of graph nodes extracted from Android apk files. Basic reporting is good. However, literature work should be added more from the year 2020-2021 Experimental design: Authors did not mentioned why they use Recursive Feature Elimination (RFE) and Fisher Score (FS) to conduct feature selection in comparison to other studies. Please explain why you used these methods. Furthermore, Is the analysis is sufficient enough by using SVM as a model., why not considering other well-known classifiers that classifies better than SVM? Please explain list of feature selection in tabular format. In table 6 it seems that increasing number of features will increase accuracy, which is normally not true. So, what will be the accuracy at 60 features? Validity of the findings: Please explain the parameter space of LightGBM in detail. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MALWARE DETECTION FRAMEWORK BASED ON GRAPH VARIATIONAL AUTOENCODER EXTRACTED EMBEDDINGS FROM API-CALL GRAPHS Review round: 1 Reviewer: 3

Basic reporting: The paper is well written and the content is well connected. In terms of basic reporting I would recommend following corrections: 1) The abstract of the paper is lengthy. Please reduce the content of the abstract and omitting detailed methodology and only report the primary results. 2) The Figure methodology diagram should be re-drawn as a block diagram. The diagram should be divided in to different connected components. 3) The contributions of the paper should be written in bullet points or with numbering. 4) The paragraph after the contributions should be re-written. 5) The API extraction from the data set is not explained. How did you extracted the API calls? Which tools did you used? This should be explained. Experimental design: The experiments are presented clearly, however, i have concern with the explanation of dataset. Please address the following: 1) Include a section which explains the dataset. I do not see the number of benign apps being used. Is the dataset balanced or not? 2) Include a table to present/explain the dataset. Validity of the findings: Satisfied. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MALWARE DETECTION FRAMEWORK BASED ON GRAPH VARIATIONAL AUTOENCODER EXTRACTED EMBEDDINGS FROM API-CALL GRAPHS Review round: 2 Reviewer: 1

Basic reporting: It is ready for publication Experimental design: It is ready for publication Validity of the findings: It is ready for publication Additional comments: It is ready for publication

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MALWARE DETECTION FRAMEWORK BASED ON GRAPH VARIATIONAL AUTOENCODER EXTRACTED EMBEDDINGS FROM API-CALL GRAPHS Review round: 2 Reviewer: 2

Basic reporting: Satisfied Experimental design: Changes are made as mentioned Validity of the findings: Satisfied Additional comments: Suggested changes are modified.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ETHICAL VISUALIZATION OF THE NORTHCOVID-19 MODEL Review round: 1 Reviewer: 1

Basic reporting: I’m impressed how many words can be spent for explaining and justifying the design of a simplified user interface. The paper reads very well, however, I can’t really assess its value as the videos themselves, not to speak of the system generating the videos on the fly, are not available, in contrast to the original website, https://covid.datalab.science, for which the simplified user interface has been developed for. I suggest to add a few words explaining how the outcome of the reported work will be used/made available in the future. Experimental design: no comment Validity of the findings: Can't be assessed as the main result of the paper (a video-generating add-on to a given web interface) is not available. Additional comments: This paper has nine authors; it could be helpful to see how each of the authors contributed to the reported work. There are no raw data. minor: line 119: there seems to be something missing after ‘section’;

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ETHICAL VISUALIZATION OF THE NORTHCOVID-19 MODEL Review round: 1 Reviewer: 2

Basic reporting: The authors examine the outputs of an epidemic model called “NorthCOVID-19” along with its web interface to see what issues it may have when being interpreted by the general public. They identified three main issues that could cause anxiety and confusion and focused on creating a solution that would thoroughly explain the results in an ethical, easy-to-understand format. They built audio and video methods including using AI to communicate the epidemic model results. This is a well-written paper that I recommend to be accepted. Experimental design: The experimental design is clear with sufficient details to reproduce the experiments and results. Validity of the findings: Conclusions are well-stated. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ETHICAL VISUALIZATION OF THE NORTHCOVID-19 MODEL Review round: 1 Reviewer: 3

Basic reporting: Overall it looks interesting. However, I am not too sure if the assumptions made are true. Experimental design: Assuming that the proposed model is for public use. I don't think so. You cannot find the model anymore. Validity of the findings: One needs to provide some facts about how much the model was used by the general public. How independent the study can be if the first author of this paper is the second author of the published model paper? A sample of 40 is not too bad, but not knowing who the participants are in real trouble. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN ETHICAL VISUALIZATION OF THE NORTHCOVID-19 MODEL Review round: 2 Reviewer: 1

Basic reporting: () re individual contributions of all authors: I assume that this will be visible in the final paper; so far this information is not included. () https://github.com/andrfish/NorthCOVID19, Video Generation Script for the COVID Crushers, introductory comment: “This script takes input from NorthCOVID-19 (see model here) and produces a video animation (see sample_format.pdf) of the results. The file "sample_output.csv" is an example of what the output will look like, "sample_parameters.json" is an example saved file of the parameters used, and "sample_results.json" is an example of the results output to the website from the simulation.” -> none of the links work - ‘NorthCOVID-19’ , ‘model’; for some of the pointers there are no links given at all, eg. sample_format.pdf, sample_results.json, sample_parameters.json. () “The result was a video” thanks for adding the video generation feature to the original website https://covid.datalab.science; I suggest to move the link to this website, hidden so far in a footnote, to a more prominent position. () “Please do not leave this page, the generation may take up to 10 minutes and the download will start automatically.” -> having to wait up to 10’ (it actually took almost as long) is a rather long delay for getting something simplified compared to what I immediately get and comprises actually more information. () the last slide in the generated “video” (is more a sequence of slides) needs some normalisation to make the results visible. in summary, I’m still amazed how many words one can spend for explaining and justifying the design of a simplified user interface. I personally would prefer the direct output as offered by the original website https://covid.datalab.science; not to say that I somehow feel offended that I - as a user- is considered not to be able to understand what the original diagrams are telling me. But I’m happy to accept that I might not belong to the actual target group of the tool. Experimental design: N/A Validity of the findings: N/A Additional comments: none

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MATHEMATICAL COVID-19 MODEL WITH VACCINATION: A CASE STUDY IN SAUDI ARABIA Review round: 1 Reviewer: 1

Basic reporting: - Experimental design: - Validity of the findings: - Additional comments: Please add contribution with bullet mark in introduction (Please see and cite [1-9]) Please add lit. review table and add your research in end of table and show gap research (Please see and cite [1-9]) Please add notation list and classify to sets (indices) , parameters, decision variables (Please see and cite [1-9]) Please add Managerial insights and practical implications (Please see and cite [1-9]). Please add results with bullet mark in conclusion (Please see and cite [1-9]). Please suggest uncertainty form like fuzzy, robust optimization, data-driven and stochastic. Please suggest risk form like CVaR, VaR. [1] Lotfi, R., Kheiri, K., Sadeghi, A., & Babaee Tirkolaee, E. (2022). An extended robust mathematical model to project the course of COVID-19 epidemic in Iran. Annals of Operations Research, 1-25. [2] Lotfi, R., Kargar, B., Gharehbaghi, A., & Weber, G. W. (2021). Viable medical waste chain network design by considering risk and robustness. Environmental Science and Pollution Research, 1-16. [3] Lotfi, R., Mardani, N., & Weber, G. W. (2021). Robust bi-level programming for renewable energy location. International Journal of Energy Research, 45(5), 7521-7534. [4] Lotfi, R., Kargar, B., Hoseini, S. H., Nazari, S., Safavi, S., & Weber, G. W. Resilience and sustainable supply chain network design by considering renewable energy. International Journal of Energy Research. [5] Lotfi, R., Mehrjerdi, Y. Z., Pishvaee, M. S., Sadeghieh, A., & Weber, G. W. (2021). A robust optimization model for sustainable and resilient closed-loop supply chain network design considering conditional value at risk. Numerical Algebra, Control & Optimization, 11(2), 221. [6] Lotfi, R., Yadegari, Z., Hosseini, S. H., Khameneh, A. H., Tirkolaee, E. B., & Weber, G. W. (2020). A robust time-cost-quality-energy-environment trade-off with resource-constrained in project management: A case study for a bridge construction project. Journal of Industrial & Management Optimization. [7] Lotfi, R., Safavi, S., Gharehbaghi, A., Ghaboulian Zare, S., Hazrati, R., & Weber, G. W. (2021). Viable Supply Chain Network Design by considering Blockchain Technology and Cryptocurrency. Mathematical Problems in Engineering, 2021. [8] Lotfi, R., Sheikhi, Z., Amra, M., AliBakhshi, M., & Weber, G. W. (2021). Robust optimization of risk-aware, resilient and sustainable closed-loop supply chain network design with Lagrange relaxation and fix-and-optimize. International Journal of Logistics Research and Applications, 1-41. [9] Lotfi, R., Kargar, B., Rajabzadeh, M., Hesabi, F., & ?zceylan, E. (2022). Hybrid Fuzzy and Data-Driven Robust Optimization for Resilience and Sustainable Health Care Supply Chain with Vendor-Managed Inventory Approach. International Journal of Fuzzy Systems, 1-16.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MATHEMATICAL COVID-19 MODEL WITH VACCINATION: A CASE STUDY IN SAUDI ARABIA Review round: 1 Reviewer: 2

Basic reporting: The English is poor. Authors have to revise the paper carefully. For example, "The discovery of a new form of corona-viruses in December 2019, SARS-CoV-2, commonly named covid-19, has transformed the world." the word transformed here is wrongly used. "The world has been waiting for the vaccine for almost one year." may be more than one year? Also line 261 and line 265 should be corrected. Experimental design: The numerical experiment are given. However, wome detail illustration should be given on these results with figures. Validity of the findings: The results in this paper deserve publication. Additional comments: The references are not completed, such as Nonlinear Dynamics 106, 1491-1507, 2021, Nonlinear Dynamics 106, 1347-1358, 2021.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MATHEMATICAL COVID-19 MODEL WITH VACCINATION: A CASE STUDY IN SAUDI ARABIA Review round: 1 Reviewer: 3

Basic reporting: The COVID-19 pandemic has already spread throughout the world and the people are aware about the diseases and they are using precautions about the pandemic. But, still the covid-19 is spreading very quickly. Formulate a mathematical model for the transmission dynamics of COVID-19. Few minor issues: ---- The abstract is a little thin and does not quite convey the vibrancy of the findings and the depth of the main conclusions. The authors should please extend this somewhat for a better first impression. ---- The manuscript lacks motivation. Author needs to include the motivation of the study. -----To stop the spread of the diseases vaccine is needed. But, in absence of the vaccine people must have maintain the social distancing. In order to maintain the social distancing must obey the modeling rule. The introduction need to be improved by incorporating some recent references of COVID-19 pandemic. To do so, I suggest some modeling work must be included in the references: "Modeling and forecasting the COVID-19 pandemic in India, Chaos, Soliton & Fract. 139 (2020) 110049", “Mathematical modeling of the COVID-19 outbreak with intervention strategies, Results in Physics, 2021, 104285”. Experimental design: Need to explain properly the parameter estimation. For parameter estimation and model validation authors should include the manuscripts: "Forecasting the daily and cumulative number of cases for the COVID-19 pandemic in India, Chaos, 30(7) (2020) 071101", “Impact of social media advertisements on the transmission dynamics of COVID-19 pandemic in India, Journal of Applied Mathematics and Computing (2021)” Validity of the findings: For model validation and proper writing of the manuscript authors should include the manuscripts: "Modeling the dynamics of COVID-19 pandemic with implementation of intervention strategies, The European Physical Journal Plus (2022)"; "Mathematical modeling and optimal intervention strategies of the COVID-19 outbreak, Nonlinear Dynamics, (2022)". Additional comments: No coemmnts.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MATHEMATICAL COVID-19 MODEL WITH VACCINATION: A CASE STUDY IN SAUDI ARABIA Review round: 2 Reviewer: 1

Basic reporting: It is suitable for publish Experimental design: It is suitable for publish Validity of the findings: It is suitable for publish Additional comments: It is suitable for publish

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MATHEMATICAL COVID-19 MODEL WITH VACCINATION: A CASE STUDY IN SAUDI ARABIA Review round: 2 Reviewer: 2

Basic reporting: Improved. Experimental design: Improved. Validity of the findings: Improved. Additional comments: None

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN APPROACH TO FILL IN MISSING DATA FROM SATELLITE IMAGERY USING DATA-INTENSIVE COMPUTING AND DINEOF Review round: 1 Reviewer: 1

Basic reporting: I support the publication of this work as the authors present a non-easy task through a creative approach. However, I suggest they follow the below suggestions before this research can be considered for publication. I also noted that the quality of the figures provided is not the best and needs to be improved. Experimental design: To improve the experimental design of filling missing (infrared/missing/clouds) satellite-data I suggested following a different approach. It relates to the use of non-missing data from model outputs and adding different percentages of clouds/missing data randomly. Better yet, this new approach can be developed on multiple time steps fields, thus you fully approach the strength of your method by following the existent nature of surface features. Thus you will be able to develop stronger statistics (figure/table) to conclude how well/bad this new approach solves the filling of missing data while ensuring that your newly created data do not create false surface features and it is consistent with the original data. It is also important that authors provide the original data and the methodology (Python/Matlab) scripts to recreate their analysis. Validity of the findings: I believe the validity of the findings is questionable based on the current approach. See above for suggestions on a new approach to better evaluate statistics and skipping the creation of false features, which is conditional to some interpolation approaches. Additional comments: I suggest you invest a little bit more time in this analysis and provide a robust analysis that we all in the satellite-data world will appreciate.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN APPROACH TO FILL IN MISSING DATA FROM SATELLITE IMAGERY USING DATA-INTENSIVE COMPUTING AND DINEOF Review round: 1 Reviewer: 2

Basic reporting: The article has a clear and unambiguous English. The English level is sufficient for an overall correct understanding of the text, taking into account that is not likely the mother tongue of any of the authors. However, I have highlighted (in my "General Comments") a few terms that should be replaced by more appropriate ones Literature references are sufficient to illustrate the scope of the research and to give adequate background. The structure of the article is correct and the authors share all the material needed for a thorough revision. The objectives of the research and the results are clearly stated Experimental design: The scope of the research fits in the aims and scope of the journal. The authors propose a methodological improvement to well established gap-filling methods for ocean colour satellite data. The experimental design is correct and the obtained processor allows to apply the DINEOF program to high-resolution data through a segmentation procedure. This is the main advance in the developed processor, together with a refinement in the merging procedure that, however, it is not clear to me how different is from the previous method, based on pixel averages (see my comment in the "General Comments" section). The advance in knowledge is very small, but the results are of interest to the community of satellite data processing. Validity of the findings: The results presented are insufficient to support the conclusions of the authors. I think this is the weakest part of the article, that needs to be improved. In my general comments below I suggest the authors different validation exercises that would serve to improve the soundness of the conclusions. In summary: - For the assessment of the differences in the merging method, artificially removing a certain number of valid pixels in the original images is proposed. These pixels would be then used for validation - Instead of the visual inspection in Figure 6, some statistical analysis is asked for Additional comments: Below, my notes with comments and amendments requests throughout the text: Abstract: Line 19: Replace “capture” by “acquisition” Line 23: Replace “For proof of concept” by “As a proof of concept” Line 26: Replace “chlorophyll level” by “chlorophyll concentration” Throughout the text, the term “acquisition” should be used preferably, instead of “capture”. Introduction: Overall, the introduction to the oceanographic background of the work needs a revision, preferably by an expert in the field, to correct several misunderstandings on the biological and physical basis of the phytoplankton role in Oceans. Line 28: Replace “plant organisms” by “photosynthetic organisms”. Besides microscopic algae, Cyanobacteria are the main constituents of phytoplankton. The references to the impact of phytoplankton are basically incorrect and should be removed or changed. The authors confound “weather” with “climate”. Moreover, the term “weather balance” is vague and meaningless. Line 31: replace “biochemical cycle” by “biogeochemical cycle” It is not clear how phytoplankton could provide information to predict impact of Climate change. The whole first paragraph needs to be revised. Line 34: The most used abbreviation for Chlorophyll-a is “Chl-a”, instead of “Chlo-a” Line 38: Replace “interaction” with “impact”. Chl-a interacts with light. Its interaction, in turn, determines the magnitude and shape of the reflectance spectra in water bodies. Line 39: Replace “closed” by “close” Line 42: replace “space” by “scene” Line 53: What does “an overall temporal record” means? Line 74: Again the confusion between “weather” and “climate”. The sentence in line 74 could be correct, because there are many meteorological satellites used in atmospheric modelling and weather forecasting. But the three references included are related to ocean color sensors and to the relationship of phytoplankton and climate Methodology Lines 89-90: The sentence has no sense: “importance of chlo-a in chlorophyll studies” ?? Line 103: What does “Python is a representative” means? Line 110: Replace “passed by” by “passed over” The formula for filling holes in the preprocessing step consists in calculating the average of the n nearest valid pixels (with data). Which is the difference in this calculation with the so-called “classical” method based on the average? The adjustment applied to reduce the impact of differences in capture conditions is not clearly explained Results In the explanation of the merging procedure, it is stated that “the ortho-mosaic with most data related to chlorophyll is chosen and tagged as base-image”. However, in the example shown in the results, the VIIRS-SNPP image was taken as the base-image despite not being the one with more valid data. Can the authors explain the reasons for that choice? Figure 5 shows that the classical method (GPT-SNAP) overestimates values with respect to the proposed method. But, why it is assumed that the proposed method is producing the correct values? A good way of proving this is to produce artificial holes (by removing valid pixels with a simulated pattern) and then use the valid original pixels to test the performance of the merging procedure. Authors are encouraged to do this experiment. Line 209: Replace “fussed” by “fused” In the comparison described lines 209 to 215 (figure 6) it is very difficult to visually observe the differences in concentration claimed by the authors. A comparison of histograms and/or statistic test on differences would be more informative than the visual inspection.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN APPROACH TO FILL IN MISSING DATA FROM SATELLITE IMAGERY USING DATA-INTENSIVE COMPUTING AND DINEOF Review round: 1 Reviewer: 3

Basic reporting: The manuscript "An approach to fill in missing data from satellite imagery using data-intensive computing" by Rivera-Caicedo et al is well structured and the scientific investigation is clearly presented, including a basic context and motivation. However, the scoping is not clear and the methodology shows little originality. The use of the DINEOF method for gap-filling is not new, and the literature contains other titles, including MODIS-based work. Data Interpolating Empirical Orthogonal Functions (DINEOF): a tool for geophysical data analyses (2011) Reconstruction of MODIS total suspended matter time series maps by DINEOF and validation with autonomous platform data (2011) Analysis of gap-free chlorophyll-a data from MODIS in Arabian Sea, reconstructed using DINEOF (2018) Exploratory Analysis of Urban Climate Using a Gap-Filled Landsat 8 Land Surface Temperature Data Set (2020) Experimental design: Although the authors claim they introduce "a new approach for filling in missing data from satellite imagery", the level of originality is low. The DINEOF method is already well-known, and the authors present some references. For me it is not clear what is new, and the authors do not emphasize the novelty. Validity of the findings: The validation of the findings is a very important part is such studies. I cannot find a distinct validation section, where the results are compared either with other methods or to measurements. This is a fundamental part which makes me have serios doubts about the quality of the work. The so-called new method is divided into modules, which in principle is good, but in fact only one module refers to data-gap filling, and the rest are basic image processing. Additional comments: I strongly recommend to justify better your results, refer to more previous work and resubmit the manuscript.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN APPROACH TO FILL IN MISSING DATA FROM SATELLITE IMAGERY USING DATA-INTENSIVE COMPUTING AND DINEOF Review round: 1 Reviewer: 4

Basic reporting: The authors present an approach to fill in missing data from satellite imagery using data-intensive computing. The approach was divided into three main modules. The idea is interesting; however, the authors must improve several details before accepting the paper. The authors must improve the approach implementation explanation. There are gaps in how it was implemented. As explained, it seems that they only divided the data into smaller sets to be processed. Also, several plots must be explained in another way. For example, in figure 5, the comparison is not understood. the authors must justify parameters used in the implementation ( for example 3 x 3 window) Table 1 is missing Experimental design: The authors present interesting experiments and results. However, they must find a metric to validate the results; a visual comparison is not enough. This metric will allow comparing the proposed approach with other methods. Also, they can report processing time a compare it with the DINEOF algorithm processing time. Validity of the findings: no comment Additional comments: The idea is a novelty; however, it needs to be clarified to understand the real impact and originality.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: AN APPROACH TO FILL IN MISSING DATA FROM SATELLITE IMAGERY USING DATA-INTENSIVE COMPUTING AND DINEOF Review round: 2 Reviewer: 1

Basic reporting: The authors present an approach to filling in missing data from satellite imagery using data-intensive computing. The approach was divided into three main modules. The idea is interesting; however, the authors must improve several details before accepting the paper. The authors must improve the approach implementation explanation a justify what is the main novelty of their method compared to the state of the art. As explained, it seems that they only divided the data into smaller sets to be processed and this is not enough. Also, a computational complexity analysis must be added to understand and clarify the contribution of the approach. Experimental design: Experimental design The authors present interesting experiments and results. However, they must find a metric to validate the results; a visual comparison is not enough. This metric will allow comparing the proposed approach with other methods reported in the literature. for example the complexity of their approach. Also, they can report processing time a compare it with the DINEOF algorithm processing time with previous work reported. Validity of the findings: no comment Additional comments: the authors must clarify the real impact and originality of the proposed method. Currently, this is not enough to accept the paper.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM-EFFECTIVE EXACT MULTIPLE PATTERNS MATCHING ALGORITHMS FOR BIOLOGICAL SEQUENCES Review round: 1 Reviewer: 1

Basic reporting: Avoid using abbreviations in abstract unless it’s a time complexity. Please provide a table for all the abbreviations. Please merge the Prior work and Important Findings and Related work section into one section. Experimental design: Theorem 1 and 2 are defined twice. It should be defined once only in the proposed methodology section. Validity of the findings: The Section- ‘Simulation detail and analysis with algorithms evaluation criteria’ should be added as a subsection within Results and Discussions. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM-EFFECTIVE EXACT MULTIPLE PATTERNS MATCHING ALGORITHMS FOR BIOLOGICAL SEQUENCES Review round: 1 Reviewer: 2

Basic reporting: The article claims to proposes an efficient quantum solutions for exact multiple pattern matching to process the biological sequences. It appears that a lot of effort has been made in writing the article for presenting the solution through complexity analysis of algorithm. Pl. refer to the attachment. Experimental design: Experimental Design is weak due to limitation of current hardware technology so cannot be basis of a publication. Further, Grover's search can overshoot if the number of solutions are not known in advance, which is going to be the case when t exact pattern matches are not known, so how are you going to handle such cases while performing simulation. Ideally there should be table for depicting numerical values of variables like N, m, t etc. along with constants C etc. and performances in time and space rather than equations. Further Table 11 and 12 should also report on the number or percentage of patterns correctly identified, incorrectly identified and incorrectly missed etc. Validity of the findings: The article claims to proposes an efficient quantum solutions for exact multiple pattern matching to process the biological sequences. It appears that a lot of effort has been made in writing the article for presenting the solution through complexity analysis of algorithm. The article claims to find all Additional comments: This work should be divided into two parts viz., theoretical and application part with simulations giving details on experimental accuracy of the proposed algorithm.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM-EFFECTIVE EXACT MULTIPLE PATTERNS MATCHING ALGORITHMS FOR BIOLOGICAL SEQUENCES Review round: 2 Reviewer: 1

Basic reporting: 1.All the previous concerns have been addressed except usage of abbreviations in abstract. 2. English should be corrected.Grammatical errors like "At present, there is no effective quantum solution exists to 26 process multiple patterns." should be corrected. 3.In "Motivation and contribution of work" a point is given as follows: Existing Experimental design: 1.In Proof of Theorem 1 what does 't' stand for. Validity of the findings: In the Results Section, observations are given Table wise, instead the findings that they emphasize should be given as headings and relevant Tables should be included under them. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: QUANTUM-EFFECTIVE EXACT MULTIPLE PATTERNS MATCHING ALGORITHMS FOR BIOLOGICAL SEQUENCES Review round: 3 Reviewer: 1

Basic reporting: All previous concerns have been addressed. A table containing all abbreviations should be added. Experimental design: The corrections have been done as per requirement. Validity of the findings: The impact and novelty has been properly presented Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: FPGA-BASED SYSTOLIC DECONVOLUTION ARCHITECTURE FOR UPSAMPLING Review round: 1 Reviewer: 1

Basic reporting: There seems to be an issue with citation through out the text, for example in line 59 "Lu et al. Lu et al. (2017)" should be "Lu et al. (2017)". This repeated referencing should be an easy fix by fixing the format file. In line 128 " (i) Interpolation techniques" and line 129 "(iii) Transpose Convolution need references." citations are needed. Experimental design: No comment. Validity of the findings: No comment. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: FPGA-BASED SYSTOLIC DECONVOLUTION ARCHITECTURE FOR UPSAMPLING Review round: 1 Reviewer: 2

Basic reporting: This paper presents a deconvolution accelerator that upsample n×n input to 2n×2n output by convolving with a k ×k kernel. The proposed architecture does not insert and pad zeros and thus removes the redundant computations to achieve high resource efficiency with reduced number of multipliers and adders. The paper is well-written and easy to follow. Adequate literature reviews has been discussed and background clearly explained. The paper structured well along with formatted plots and figures. The result section presents a good proof of the established hypothesis. Experimental design: The article has articulated well around the defined research question. The results are reproducible and well presented. Validity of the findings: The results have been evaluated against benchmarks, the dataset has been explained and conclusions are stated clearly along with highlighted contributions. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: FPGA-BASED SYSTOLIC DECONVOLUTION ARCHITECTURE FOR UPSAMPLING Review round: 1 Reviewer: 3

Basic reporting: Abstract: * Please clearly specify the precision used for operations like multiply/add etc in the proposed architecture. * Please provide SNR reported by prior works as a point of reference. Further, please specify how the SNR translate to increased accuracy compared to prior work? * Please use IEEE double precision instead of MATLAB double precision, if applicable. * Compared to prior work, is the performance higher or lower? Introduction: * Line 26-29: It is unclear why the citation [Daoud and Bayoumi, 2019] is used here. * Line 45-98: Please fix the citations. There are two authors cited for the same work. Example Line 48-49. * Line 59: Please clarify what is segmentation cost and how it is different from throughput. Is the segmentation cost the energy consumption or latency or something else? * Line 73: Yazdanbaksh et al has been cited but not explained. * Line 71: Only suitable for small kernels — why? Is it due to arithmetic/compute complexity or memory consumption. * Further, the metric used by the paper for evaluating the error The reported 3.6 GOps at 200 MHz would translate to just 18 Ops/cycle. Does the proposed architecture fully utilize the resources in the FPGA? Please comment on how the proposed architecture would support neural networks where multiple different deconvolutions layers with different kernel sizes and input/output sizes are used. Experimental design: The experiments focus on deconvolution layers of fixed sizes (for example, 256x256 output and 3x3 kernel). It is unclear if the evaluated sizes for inputs/output/kernels/strides are employed in real-world workloads like UNets. The numbers in Table 4 and Table 5 are generated from three sets of images - randomized, camera man, and natural. In contrast, prior works report end-to-end neural network accuracy degradation on widely used neural networks and datasets. As such, it is difficult to fully understand how good/bad the reported SNR is. To rectify this, please evaluate the impact on accuracy when using the proposed architecture for deconvolution/upsampling layers and a traditional CPU/GPU for the rest in widely used neural networks (like UNET) and datasets. Please clearly report the utilization of resources for the proposed architecture. The reported 3.6 GOps and 0.135 GOps/DSP would translate to just 27 DSPs. In this case, the proposed architecture does not fully utilize all the DSPs available on the FPGA zc7020. Validity of the findings: The comparison in Table 10 is misleading since the proposed architecture only supports deconvolution layers of a neural network, while the cited prior works (Di et al, 2020) report end-to-end performance and energy/power efficiency. Please clarify in the table, perhaps by adding a row stating whether the evaluated architectures support layers other than deconvolution. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: INVESTIGATING THE IMPACT OF VULNERABILITY DATASETS ON DEEP LEARNING-BASED VULNERABILITY DETECTORS Review round: 1 Reviewer: 1

Basic reporting: Paper summary: This paper studies the impact of datasets' distribution on the performance of deep learning-based vulnerability detectors, Three different aspects of vulnerability datasets were investigated, including the dataset granularity (function level or slice level), similarity (inter-class similarity and intra-class similarity), and the code features, such as AvgCyclmotic, AveEssential. In the evaluation part, several datasets and deep learning models are evaluated according to the three aspects, and associated insights were proposed. Advantages: + I think this paper has very good research insights on the deep learning model datasets, which does not draw much attention in previous literature. Meanwhile, the distribution of datasets is in fact very important to deep learning models' performance. + The overall experiment design is reasonable and clear. + The datasets used in this paper as well as the deep learning models are very new. Weakness: I have several concerns about this paper, as summarized below: + Overall, the writing of this paper requires improvement. The high-level structures of this paper are relatively clear, but there are flaws: 1) The abstract needs to be more succinct, especially on the method description. Such as "The training set is used to train the DL-based vulnerability detector". Such statements do not need to be in the abstract. 2) In the Experimental Result Section, I think it will be better to switch the order of Table 4 and Table 5, and their introduction. It is more intuitive to present the representation similarity of the datasets first before giving the corresponding evaluation results. 3) In Table 2, I think it will be better to add a column to show the category of the dataset, such as synthesized dataset, manually modified, and open-source software. I understand such information can be found in RELATED WORK, doing so will make the table clear and release the unnecessary burden of readers. 4) Typos such as "Dulnerability Datasets" IN Section EXPERIMENTAL SETUP Line 186. 5) In the INTRODUCTION section, redundant information on "what's deep learning" is given. I think this part could also be more succinct since the knowledge should be already known to the journal readers. + Another concern is the motivation to select the three aspects (granularity, similarity, code features) is not clearly presented in the paper. More evidence and motivation should be added to show why these three factors are important for evaluations. In addition, the similarity and code features are in fact characteristics of the raw code datasets, while granularity can be considered as a kind of preprocessing. I think it will be better to discuss them separately. + The details of experimental implementation are not given. For example, I was expecting the introduction to experiment platforms, the libraries used, experiment data such as running time. + Still in the evaluation part, the sizes of the datasets are not given. Also, the parameters (m, n) mentioned in STEP I (Line 146) are not given and not evaluated. Experimental design: As shown above. Validity of the findings: As shown above. Additional comments: no comment

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: INVESTIGATING THE IMPACT OF VULNERABILITY DATASETS ON DEEP LEARNING-BASED VULNERABILITY DETECTORS Review round: 1 Reviewer: 2

Basic reporting: This paper analyzes the inner connection between DL-based vulnerability detectors and datasets. The paper mainly focuses on the following aspects: fine-grained samples, datasets with lower inner-class similarity, and datasets with higher inter-class similarity and lower intra-class similarity. Although many other points can be explored in this kind of research, this paper grants a good view of the relationship between dataset and detector schemes. Also, this is interesting research, giving guidelines to the related research. Experimental design: This is the most important part of the paper since the author claims their contributions focus on the evaluation. More effort should be put into this part. Validity of the findings: The author needs to put more effort into the evaluation parts, as mentioned above. Some suggestions: 1. using recall rate and an F score is good, but ROC curve, EER, and accuracy are also necessary to evaluate various schemes. Since this part is the paper's focus, authors are suggested to provide detailed evaluation results in the paper. 2. How did you divide the dataset? What is the training-testing ratio? This will also impact the detectors' performance. 3. What was information loss when you applied PCA to the dataset? and what is the energy and time consumption when you remove the PCA? 4. Instead of using intra- or extra-class similarity, authors are suggested to evaluate the entropy difference between data samples. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: INVESTIGATING THE IMPACT OF VULNERABILITY DATASETS ON DEEP LEARNING-BASED VULNERABILITY DETECTORS Review round: 1 Reviewer: 3

Basic reporting: 1- This paper has not reached to the acceptable level for publication because of lacks originality and novelty. 2- The spell-checks, grammatical and writing style errors of the paper must be improved. Experimental design: No comment Validity of the findings: No comment Additional comments: 1- For readers to quickly catch the contribution in this work, it would be better to highlight major difficulties and challenges, and your original achievements to overcome them, in a clearer way in abstract and introduction. 2- Some references are too old and please add at least five references within the past one year for related work section. 3-There are unsatisfactory organization and writing in the "Conclusion" section. The authors must rewrite and reorganize this section contents. 4- The highlights are not stressed in the paper, and the innovation of the paper weakness. 5- There are some mistakes in the style of English writing in the text which are to be revised/corrected carefully.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: INVESTIGATING THE IMPACT OF VULNERABILITY DATASETS ON DEEP LEARNING-BASED VULNERABILITY DETECTORS Review round: 2 Reviewer: 1

Basic reporting: The authors solved the issues mentioned in the review comments well. This version looks good to me. Experimental design: See above. Validity of the findings: See above. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: IRIMAGE: OPEN SOURCE SOFTWARE FOR PROCESSING IMAGES FROM INFRARED THERMAL CAMERAS Review round: 1 Reviewer: 1

Basic reporting: This contribution is about processing of thermal images. In general, this is a complex problem, users usually want to determine the temperature of the object, but thermal cameras measure the thermal radiation of the thin outer layer of the object. The English is ok, but I am not native speaker, and I am not relevant to evaluate English. Literature review is sufficient, the article structure is appropriate. As I am first time as reviewer in this journal, it is a little bit difficult for me to find additional information. Experimental design: The research should be supported by a physical overview. The reader appreciated that. You write: the theoretical background of the algorithm used in IRimage and how it was implemented is included in this paper as Article S1 First time I downloaded only the main text.only and I didn’t find more information about physical overview. I don’t find in this paper Article S1 (I find it later), why it is divided on more parts? Maybe it's complicated from the publisher This article (main text) is basic introduction to your system (software) with case projects, but it seems so inappropriate to me as a scientific paper for journal. Only some sentences and reference to S1.are In material and methods, and so the article looks just like a guide to the software and usage examples. However, this research is certainly interesting, usable, and valuable. Validity of the findings: The article presents usable software for processing thermal images. Cheap thermal cameras only output to jpg and do not give raw data, this is an example of thermal cameras on a DJI Mavic drone, for example, which complicates scientific use. Additional comments: General comments: this article summarizes the state of the art of using low-cost thermal cameras and gives an opportunity to process data from these instruments based on scientific research. Research and information are useful for practice. I recommend the article for publication. But I have a more fundamental note to the overall text: why is the theoretical part outside the article? I would like to at least welcome the extension to essential physical information and fundaments in the Materials and methods section.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: IRIMAGE: OPEN SOURCE SOFTWARE FOR PROCESSING IMAGES FROM INFRARED THERMAL CAMERAS Review round: 1 Reviewer: 2

Basic reporting: This paper describes the functionality of the open source software IRimage which can be used for processing raw thermal images taken with several cameras of one brand FLIR and can potentially be used for other brand cameras. Some example results are presented on 14 thermal images obtained from a (non-verified) public source of thermal images ( Wikimedia ). In addition, an example use case of monitoring the temperature of plant leaves is presented of well and non-well hydrated plants during 24 cycles where many parameters vary in time. The paper is clearly written in professional English language. The introduction give a good background in support of the need and usability for the software IRimage presented. Experimental design: The author states that in this paper the software is validated against standard software. The reviewer does not agree that the method used can be considered as scientifically validation procedure using some sample images of a non-verified source and presenting the results in graphs at large scales (e.g. figure 2) were a difference of even 5 degree cannot be distinguished. In the example use case, too many variables have an influence on the measured/calculated temperature to prove the accuracy of the software. The details from the case presented distract from the purpose of this paper to present the IRimage software and could be a paper by itself. For a true validation process, the thermal images should be taken with different cameras of the same object under controlled conditions. Not having all these camera available, the reviewer understand that this is a difficult task. However, with only 3 different cameras, the author would already present the validity of the software. As example case, it would be more convincing to have only 1 or 2 variables for the temperature changes in time compared to a temperature reference e.g. a black body source temperature controlled by a current (temperature phantom). The case with the leaf temperature presented is nice as example for the many processing possibilities but cannot be claimed as a confirmation for accuracy. Validity of the findings: If the author would adapt the text by stating some examples are presented of processing some random thermal images using IRimage it would be acceptable. Still the temperature comparison graphs like in figure 2 need to be presented zoomed in on the temperature scale of e.g. 10 degree so the accuracy within 0.1 degree can be distinguished. Additional comments: The reviewer is convinced the software IRimage will be well received by the scientific community as a useful tool to process raw data from thermal cameras as long as this data is accessible. It would be great if this would also be used for small thermal imagers like the FLIR-One. This paper gives a good introduction in the potentials of the software. However, a scientific validation process is needed in which other researcher could contribute. The paper should be adapted not claiming the software has been validated by the examples presented. The example case could be presented with far less detail or replaced by a more simple case as suggested.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: IRIMAGE: OPEN SOURCE SOFTWARE FOR PROCESSING IMAGES FROM INFRARED THERMAL CAMERAS Review round: 2 Reviewer: 1

Basic reporting: The reply of the author and revisions in the manuscript are adequate and meet the expectations of the reviewer. The 'claim' that the software is validated have been removed and replaced by 'compared to existing software'. It would still be preferred to have the software validated in a controlled setting with various thermal cameras (in the future paper). Now there is no claim of validation, the case presented is a nice illustration but still a more simple case would be preferred with less variables that can influence the temperature. Experimental design: In this revision more theoretical background is given on the algorithm and parameters used to determine the object temperature. This greatly improves the scientific basis for the paper. No additional comments compared to the first version Validity of the findings: The presentation of the thermal images in figure 2 extended with a detailed section, as suggested by the reviewer, enables a better comparison (and convincing) between IRimage and FLIR Tools down to 0.1 degrees. The example use case is presented in the results extensively and could be shortened e.g. leaving figure 6 out. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: EVALUATION OF FEATURE PROJECTION TECHNIQUES IN OBJECT GRASP CLASSIFICATION USING ELECTROMYOGRAM SIGNALS FROM DIFFERENT LIMB POSITIONS Review round: 1 Reviewer: 1

Basic reporting: The paper is concerned with feature projection methods applied to the domain of EMG signal processing. In particular, data from several sensors is collected in order to classify grasps. Since the number of computed features may be large in such tasks, it is desired to reduce the dimensions before the application of classification algorithms. The authors test several methods for dimension reduction, some linear and some non-linear, with an emphasis on the non-linear technique named spectral regression extreme learning machine. This method was shown to have the best performance. The level of English is good. The background and Introduction are adequate. Figures are of good quality. Experimental design: The experimental designs is explained in detail, the presented figures are helpful. Some comments regarding the algorithmic implementation are detailed bellow. Validity of the findings: Experimental results are adequate, and the findings are solid. Additional comments: Overall, the topic is of interest, and the paper is solid, but I would say that the contribution mainly focuses on the question of comparison between dimension reduction techniques and a focus on the dataset. The fact that limb position changes in the dataset, is more challenging compared to some previous studies (still this seems like a not-very-hard classification task). Comments: Page 3, line 126 – This is more challenge  more challenging Feature projection section: • It is mentioned that the number of coordinates in the projected space is the number of grasp types minus one. It may be the case that more dimensions are needed for a better separation. Did you try other numbers? Page 6, line 191 and on: • The train-test separation was not clear to me. I assumed that the classification is not multi-subject, rather it is single subject (train-test belongs to one subject). Is that so? • Does the train data contain all of the recorded positions? Or are some 'new' positions left out for the test data? • It would be interesting to see the performance of the methods as the number of training samples is reduced. • Please explain in more detail how the projected coordinates in each method are extended to the test data. In PCA, new points are normalized and projected to the train model. How about tSNE? It is not trivial to extend these coordinates. Is the model built over with the test data? • Same question for SRELM. Page 6 line 217: SRELM • Since this method is new and probably most readers are not familiar with it, it would be useful to add a more detailed explanations of its construction, maybe in a pseudo code manner (instead of sending the reader to past papers). Results section • Please add an opening sentence before the sentence in line 250 (Page 8) that starts with "In Table 2…"

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: EVALUATION OF FEATURE PROJECTION TECHNIQUES IN OBJECT GRASP CLASSIFICATION USING ELECTROMYOGRAM SIGNALS FROM DIFFERENT LIMB POSITIONS Review round: 1 Reviewer: 2

Basic reporting: 1. Subheadings must be bold, followed by a period, and start a new paragraph e.g. Background. The background section text goes here... 2. Figures are of poor resolution and clarity. They need to be revised with high resolution Experimental design: 1. I suggest that you improve the description about % split of train and test set at lines 191- 192. Also give system specifications. 2. Feature extraction methods should be improved with more elaboration and illustrations. 3. More clarity on the Classifier parameter selection should be included Validity of the findings: 1. No new hybrid algorithm/enhancement based on feature projection method is proposed. 2. No simulation/Pseudo-codes or algorithm is given. 3. Novelty of work should be strongly emphasized by authors. Additional comments: 1. Intro & background to show context are well explained. Problem statement is well defined 2. Literature well referenced & relevant 3. Rigorous investigation performed to a high technical & ethical standard. 4. Conclusions are well stated, linked to original research question & limited to supporting results

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: EVALUATION OF FEATURE PROJECTION TECHNIQUES IN OBJECT GRASP CLASSIFICATION USING ELECTROMYOGRAM SIGNALS FROM DIFFERENT LIMB POSITIONS Review round: 1 Reviewer: 3

Basic reporting: No Comment Experimental design: NO Comment Validity of the findings: The authors have mentioned in their paper "In this study, the number of hidden nodes 1000 and alpha 1 223 were used. Both were determined by trial and error." It would be better if the authors gave a range or reason for picking up a certain range and then coming to the experimental value. Additional comments: 1. The paper is written well and results show a varied range of Classification accuracy. It would be interesting to comment on the reason for sometimes very low accuracy /high accuracy. Correlating the results with data distribution can help in this regard. 2. work with subjects needing prosthesis can also be a step for future work 3. The authors need to develop their own techniques / modify the existing for novelty in the paper.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: EVALUATION OF FEATURE PROJECTION TECHNIQUES IN OBJECT GRASP CLASSIFICATION USING ELECTROMYOGRAM SIGNALS FROM DIFFERENT LIMB POSITIONS Review round: 2 Reviewer: 1

Basic reporting: The paper was majorly improved. All the point I raised were answered and addressed. More experiments were preformed, especially these emphasize the influence of the train set size. Additional text was added to explain various points that were not clear in the initial manuscript. Experimental design: Adequate Validity of the findings: Adequate Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: EVALUATION OF FEATURE PROJECTION TECHNIQUES IN OBJECT GRASP CLASSIFICATION USING ELECTROMYOGRAM SIGNALS FROM DIFFERENT LIMB POSITIONS Review round: 2 Reviewer: 2

Basic reporting: No comment Experimental design: No comment Validity of the findings: No comment Additional comments: Authors have revised the manuscript as per guidelines. The proposed algorithms are added in the paper. The result section is very well written

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: EVALUATION OF FEATURE PROJECTION TECHNIQUES IN OBJECT GRASP CLASSIFICATION USING ELECTROMYOGRAM SIGNALS FROM DIFFERENT LIMB POSITIONS Review round: 2 Reviewer: 3

Basic reporting: no comment Experimental design: no comment Validity of the findings: no comment Additional comments: The paper includes all the suggestions given earlier. Clarity and Novelty aspects are covered in depth.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A SURVEY OF THE STATE OF THE PRACTICE FOR RESEARCH SOFTWARE IN THE UNITED STATES Review round: 1 Reviewer: 1

Basic reporting: The text is well written, but contains a few syntax and grammar errors (lines 169, 178, 510, 661). In line 316 the sentence structure seems to be off for me. While I (as a non native speaker) cannot determine if it is wrong, it could be improved for better understanding. The paper is very well structured and figures and tables are included in appropriate places to ease the perceiving of the data, but unfortunately the figures have some drawbacks. Figure 1 is missing an explanation for the two sets of columns. Also, it is rather unreadable due to the subfigures being too small. Removing at least the vertical dotted lines might further increase readability. Nearly all figures show absolutes. Especially with the difference regarding the absolute number of answers, I would prefer a presentation as relatives (percentages) to allow a better comparison of the answers in figures itself (i.e. figure 5 and 6) and between figures. My suggestion is to state the absolutes in the figure caption or the x-axis captions since it is an important information. In the text the authors already use mainly the percentages, referring to the absolutes in brackets. Experimental design: In line 388 it is stated that the data was sanitized, but not explained in what way. A short description of this process would be helpful. Validity of the findings: The results are US and UK centric, which is due to the nature of the survey distribution and focus and acknowledged by the authors. While I see no issue with this and most findings might be applicable for other regions, I would recommend to mention it earlier in the paper, perhaps even the abstract. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A SURVEY OF THE STATE OF THE PRACTICE FOR RESEARCH SOFTWARE IN THE UNITED STATES Review round: 1 Reviewer: 2

Basic reporting: The basic reporting is done well. I indicated on the attached pdf a few spots where the sentences are phrased awkwardly, or where there are some typos, but for the most part the English is quite good. The literature review is particularly well done. The figures were nicely done, except for a few points that I highlighted on the marked up pdf. The paper is self-contained. The results are given for the research questions. The section with the title Discussion should probably have the section title changed to Conclusions, since the Discussion is really what came in the previous section. Also, there isn't currently a Conclusions section. Experimental design: The research questions are well defined, except for one where the wording is awkward. The one with the awkward wording is highlighted in the marked up version. The investigation is well performed and the ethical aspect is explained. It is also nice to see that the full questions and data are released as open-source. It is good to see authors practicing what they preach (or eating their own dog food). :-) Enough information is given to replicate the experimental design. Validity of the findings: For the most part the findings seem reasonable. However, there are a few questionable points, as highlighted in the marked up version. A few specific comments include: 1. A bigger deal seems to be made of the differences in Figure 1 than the data justifies. The ideal time and the spent time look similar in all cases. For the cases where the paper claims the difference is noteworthy, it would be good to quantify the relative difference between the two measures. 2. A section on threats to validity is missing. One threat to validity is the respondents not knowing the meaning of the terminology, such as regression testing, requirements, software carpentry, maintaining and sustaining, design, .... Another threat is respondents self-reporting on their activities, such as what documentation they write. 3. There is no section for future work. For instance, the respondents self-report on what documentation they write, but this documentation could be measured directly by mining existing software repositories. Also, extended interviews could be held with some researchers/developers to get more qualitative data. 4. At line 643 it is stated that the most difficult aspects of software development are not technical. However, this is misinterpreting the study question. The respondents weren't asked what they found most difficult, but what aspects of software development are more difficult than they should be. The respondents might feel like the technical aspects are difficult, but not more difficult than they should be. Additional comments: The paper shows that domain data was collected, but this data does not appear to be used. Did the responses differ between the different domains? If this was not investigated, possibly because of insufficient data, this should be explicitly stated.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A SURVEY OF THE STATE OF THE PRACTICE FOR RESEARCH SOFTWARE IN THE UNITED STATES Review round: 1 Reviewer: 3

Basic reporting: The paper is very well-written. The language used throughout is very clear and unambiguous and the research is very well motivated. The introduction is really great and motivates the problem well. The tables and figures are also clear and easy to understand. The background is extensive, however the section is a little unorganized. I appreciate the list of prior surveys and Table I is very helpful, however it would be useful to have more specific details about how these prior studies differ from the current work. Additionally, the research questions (more on this later) do not seem to fit in this section. They have some background from the prior work, but are not well motivated for the overall work. There should be more background included to motivate each RQ and explain why it's included or they should be placed in a separate section (i.e. with the Methods). Overall this research is exciting and provides a valuable contribution to the field. Experimental design: The research is within the scope of the journal. The investigation and methods are rigorous and meet a technical and ethical standard. The main limitation are the research questions. There is a wide breadth between the RQs and, while the researchers show how these topics fit into the previous surveys, in my opinion the authors fail to motivate why these topics are important for this study and show how each RQ relates to each other and contributes to the overall goal of the work. Each RQ on its own seems very meaningful and relevant, however it is unclear how these connect to each other. To improve upon this, I would suggest adding to each RQ to explain why it is necessary, how it was selected, and how it fits into the current work. I think also combining the RQs with the specific parts of the survey mentioned in the Methods section would also be helpful. Additionally, the RQs lack background for some of specifics studied, especially for the SE-related questions. For example, requirements, design, and testing as software engineering processes, however other potential processes considered part of SE such as implementation, deployment, maintenance, etc. are missing in RQ1. Also the *best* practices (continuous integration, coding standards, arch/design, requirements, peer code review) have some missing options (i.e. pair programming, static analysis tools,) and would be interesting to know the details on how these were selected. Validity of the findings: The work is novel and impactful for improving how research software is developed. The findings are valid and sound based on the methods and RQs. It would be beneficial to see statistical tests used to further analyze the results and provide their significance. The organization of the results could also be improved by providing a clear answer (i.e. highlighted) to each of the research questions based on the findings. The paper is also missing a threats to validity section to explain limitations of the work (i.e. the huge discrepancy between the respondent types in the results) and how the researchers mitigated these threats. In the conclusion, I would prefer more discussion about implications of these problems and solutions for researchers and developers to improve the state of research software development for each RQ topic, especially for improving the software engineering practices. Additionally, it would be interesting to note future opportunities for this work, such as mining GitHub repos to programmatically determine the tools and practices used for research software. Overall this is a very interesting research paper with implications for improving research software, development practices, and research overall. Looking forward to seeing what's next! Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A SURVEY OF THE STATE OF THE PRACTICE FOR RESEARCH SOFTWARE IN THE UNITED STATES Review round: 1 Reviewer: 4

Basic reporting: The text follows established, scientific form and quality. I would even go as far and say that especially the text up until the "Analysis" section is written well above average quality. However, starting with the "Analysis" section, it feels like the publication was written by entirely different authors. This (long) section contains mostly tables, figures, and just a little text to describe the former. The tables and figures could be improved in several places (see detailed notes). Especially for the figures I would suggest to rethink again what they are supposed to show and whether their respective current format is really the best way to display that. Detailed comments can be found later in the review. I am not sure if this is an artifact of the review-pdf compilation, but almost all citations miss their dois and some citations are not meaningful without either a doi or an URL. Details can be found later in the review. Up until line 247 (~1/3 of the paper), the text is written in a country-neutral way (including title and abstract), leaving the reader under the impression that the whole RSE landscape is under review. However, as the authors mention themselves: "results differ across the world". Thus, I strongly suggest to mention either in the title or at at least in the abstract the limitation that this survey primarily covers only the USA. Experimental design: The raw answer data is provided, along with a (pdf) description of the questions. Something I miss is documentation of the mentioned partial dependencies of questions on answers of previous questions. This unnecessarily complicates a repetition or comparison with other surveys. In addition to providing these inter-dependencies in the supplementary material, it would be beneficial to the reader of the pdf-part of the publication if those could be highlighted or marked in some way. What I miss most is the source code of the scripts used to generate the results from the raw data, including the same for generating the figures. While it is not difficult to do this with the provided csv data from scratch, the vast number of reported numbers and plots would make scripts to generate all of those valuable, and in case of discrepancies between own and reported results the reader would be left without clear way to see where the differences are, in short: they would be very valuable to replicate the results from the data. This is especially disappointing given the connection of this issue with the topic of the paper and the background of the authors. It is entirely possible that I missed those scripts, but I could not find them either in the files submitted to the journal, nor in the data published on Zenodo. Validity of the findings: All raw data used in the report has been provided. Repetition would be hindered by the lack of dependency information and scripts for the analysis of the raw data, as mentioned in review parts 1 and 2. Conclusions are overall well stated, with only few exceptions that are mentioned in the detailed review. Additional comments: While reading the article, I collected notes of everything I spotted, even small issues like typos and the like. In the following, those notes are given in roughly the order of appearance in the text. This also means that rather more important issues are mixed with rather trivial ones . I hope this order helps to speed up the improvement of this article, because I believe it to be valuable to the scientific community. - line 80: "hey often focus on small groups or in laboratory settings" → something is off here. Maybe remove the "in" - line 84: "as briefly described next in the next subsection" → drop the first "next" - line 145: "Software Saved International Survey" is not the proper name of the survey, but rather generated from the github project name. It is commonly known as and should be better named "SSI international RSE survey". - line 169: missing parentheses around Philippe et al. (2019) - line 169: double-parentheses around Pinto et al., 2018; Hannay et al., 2009 - line 178: there is a space missing in between "integration(Shahin" - line 199: "(Nangia and Katz, 2017)": the same publication was already cited within the same sentence context. I suggest to drop the second reference. - line 201: The sentence "The AlNoamany and Borghi (2018) survey reported similar results" seems to first suggest it reports similar results as the previous sentence, which mentions a gender gap. However, reading further, the reference "similar results" seems to refer to the gender-neutral percentage of training which was discussed further up. While not technically wrong, I suggest to reword this a little to make it easier to understand the reference. - line 207: This sentence contradicts the sentence before directly, in which it is stated that only a minority of researchers (~20%) found "formal training to be important or very important". If the authors intended the meaning that this result may be caused by a lack of said training ("They don't put importance into something because they don't know it"), this should be made clearer. - line 276: "Therefore" seems to suggest to refer to the previous sentence: "The results showed that...", but this does not make sense here. It would make more sense to refer to the much previously mentioned "most of the previous surveys did not address the topic of credit": please rephrase to make this clearer. - line 279: "asking individuals who, ..., directly have important information" → "asking individuals who, ..., directly have the relevant information" - line 290: "conferences papers" → "conference papers" - line 292: "(e.g. hiring" → "(e.g., hiring" (comma) - line 306: "as is occurs" → "as occurs" - line 330: To the reader, it is not immediately clear that the statement about politeness is also a result of the earlier-quoted work by Ortu et al. (which it is). A solution could be "However, they also show that this demographic ..." - line 366: "topics" → "topic" - line 401: given that developers are also focus of later plots it would be good to also mention their number here, even if it can be calculated at this point. - table 2: numbers in "Administrative & business studies" seem to be mixed-up ("10 0" in one cell, nothing in the last cell, and indeed there seems to be an "&" missing in the source file) - table 2: minor: right-aligning the numbers would look a little nicer - figure 1: It is not clear which group is which in each sub-plot (left/right vs. combination/developers?). It is also not mentioned in the lines that describe the plot: lines 435-438. In general, the figure also suffers from the small numbers displayed in each already small sub-plot. I wonder whether it would be better to switch axes and put all 10 sub-plots as rows into one plot which can then run the entire text width. Also (but hard to see and judge as it is right now), it might be worth noting that at least the left group (combination?) wishes for more training time than they actually use. - caption figure 1: Why capitalize each word? - line 442: "The only one that is technical is testing." It might be worth noting that this is only a particular high result for the 'combination' group, but not the 'developers' group. On the other hand, for those, "requirements engineering" is comparatively high, while for the 'combination' group it is not. - lines 444-446: if possible during type-setting the paper for paper-format, it would be nice to use the space of these two lines for the figures and put the content of the lines on the next page. - figure 2: It would be better to plot the two differently-colored bars side-by-side (like in figure 3). This way, the two groups can be more easily compared between bars of the same color. If, on the other hand, the differently-colored bars are meant to be "behind each other", with the red always being "in front of" the yellow bar, this should be indicated. Even better would be a percentage-based graph, which would also make it possible to compare the two respondent groups with each other, given their different size. Also, with ~50 being the highest number seen here for groups of a size a factor of 10 larger, I wonder whether this was a multiple-choice-question or a single choice question. The difference matters, as with a multiple-choice question a vast majority effectively answered for every aspect that it is not more difficult than it needs to be, while for a single-choice question only the "most annoying" aspect could be selected, effectively producing a potentially drastically different plot than when asking the multiple-choice version. - line 444: "Focusing on the one technical aspect that respondents perceived to be more difficult than it should be": this would imply that "testing" would be the only technical aspect where respondents answered that they would be more difficult than they would be. However, figure 2 shows a lot of higher-than-zero answers for also other technical topics. "Testing" is the one technical aspect that was selected most often when it comes to being more difficult than it should be, but not the only one. - figure 3: It is not clear which group forms the basis for the plot (combination or developer or the sum of both). The numbers in the text give indication that it is likely either combination or combination + developer, but it is not explicitly mentioned. - line 450: Within the first sentence, it is again not clear which group was takes as basis for those numbers. Given that this was already an issue for more than one question, it might be a good idea to add a general indicator in the paper for each question that marks which of the groups received that particular question and that numbers without group statement always mean "everyone who received that particular question". - line 467: "we see a different story": not necessarily. The given list of topics to document might not contain, or appear not to contain, the type of documentation applied when documenting code in-line. For instance, a comment on a particular loop structure might not appear to be within "software design" it respondents correspond that answer with the general design of the overall software package. - figure 4: The text about the figure concentrates around percentages (e.g.: "Less than 30% of the respondents reported"). Therefore, changing the figure to use percentages would help the reader. Although the figure itself will show large differences, the y-axis will. Also, since the text combines 'extremely supported" and "very supported", it would be helpful to visualize those separately from the others, either by using similar colors (and dissimilar to the others), or to add a small black bar between these two and the rest. In addition, if the current order of the topics does not have special meaning, it might be good to sort them by, e.g., the percentage of combined "extremely supported" and "very supported". - line 484: From the text it is not clear which group received that question, i.e., which concrete dependency was connected to this question. - line 488: The "However" might not be warranted. It does imply that an overlap of respondents who answered "use git most of the time" and "use copy/zip most of the time" would be noteworthy and possibly troublesome. An alternative explanation might be that those with overlap indeed use both, to achieve an even higher level of backup than one single practice alone. - line 491: "The lack of use of standard version control methods": The authors state earlier, that "83/87 of respondents answered to use version control. I cannot see a large lack of use of standard version control methods given those numbers. - line 500: It would be interesting to look at the reason for the (small) difference. I could imaging that different gender ratios in specific disciplines and different availability of training in those disciplines could have an effect of comparable size. While the text does not explicitly state a causation, it might be good to explicitly state that this is indeed by itself no indication for causation. - line 510: double "half half": the first likely should be a "than" - lines 511/514: like at line 500 it would be interesting (and should be within the data collected) to see whether this could be a correlation with something else, like differences in availability of training and gender ratio differences in different disciplines. - figure 5: plotting percentages would allow comparisons between the different groups, while plotting absolute numbers do not really add benefit. - figure 8: percentages would be a better choice to show the data. - figure 9: same as for figure 8, but this is even more important here, where the current plot does not really enable a good comparison between the two groups shown here. - figure 10: also here, a percentage plot would be more useful than absolute numbers - line 563: The text states "Figure 10 shows, the respondents saw little chance for career advancement for those whose primary job is software development". However, when looking at figure 10, the group of "developers" show a far better opportunity for advancement (50/50 no/yes) when compared to the other two groups (far lower numbers of "yes" compared to "no"). Also, it is interesting to see for the group of researchers only, the group of "don't know" is the largest overall. - line 565: with so big differences between groups in figure 10, I wonder how relevant the statement concerning gender is really for the gender topic and how much could result from different gender ratios in the groups plotted here. - figure 11: percentages would be a little more useful. Luckily, the total numbers are somewhat similar here. - figures 12&13: percentages would be really important here, as the absolute numbers for developers and the combination group are so different that gets really hard to see the composition of the numbers for developers. - caption of figure 12: "sfaff" → "staff" - line 597: I would argue that availability of resources is about as important then the other three mentioned here, for the combination group reaching 30 for 'extremely' and 'very' important combined, and even if hard to see, this seems to be even more important for the group of developers. Thus, I feel it should be mentioned as important as well. - caption of figure 15: "current" → "currently" - caption of figure 16: "Does" → "do", "casess" → "cases" - figure 18: percentages would help comparing the different groups with each other. - line 634: "This answer again seems at odds"...: Not necessarily. Assuming most of these projects are within scientific institutions, those more likely do have one of both of these, so while the project might not have any, it would be covered by the institutional one (but the question only asked about the project specifically). Also, there are different views of the effectiveness of these measures, which might lead projects to adopt different practices to further inclusion. - figures 19, 20 and 21 are for some reason displayed differently than most of the others before: here, the groups use different colors and the answers are displayed as different bars, while it is the other way around for the figures beforehand. It would be easier for the reader if this would be done more consistently. Besides consistency, using percentages instead of absolute numbers (and swapping variables as mentioned above) would the reader judge differences between different groups a lot easier. - line 653: I would argue that any time spent on debugging is too much, as ideally there would not be any need for debugging. In this sense, it is not surprising that a lot of people responded that they spend more time for debugging than they ideally would (because that would be no time at all). - line 661: I suggest to add an "of" inside of "the development *of* research software" - line 721 should likely be cited as specified on F1000: https://f1000research.com/articles/9-295 - line 724: Is there a DOI or an URL available for reference? - line 725: There seem to be two (different) dashes in between the page numbers. - line 726: Likely, at least the word "interdependencies" should start with a capital "I" - line 728: URSSI should be all capital; also a link to the data or a doi would be nice. - line 744: UK should be all capital - line 793: "investigating" likely should start capitalized - line 797-799: The doi to the data is missing: 10.5281/zenodo.2585783. Without the doi, the citation is hard to use: a reader would not even know that the name refers to a github organization without utilizing a search engine. Also, a link to an analysis of the relevant data would be interesting to readers: https://www.software.ac.uk/blog/2018-03-12-what-do-we-know-about-rses-results-our-international-surveys - line 800: "an" should likely start capitalized - line 811: That citation should at least include the link to the blog entry. As it is, one would have to rely on a search engine to try to find what is meant. - line 819: There seem to be two dashes: "CIDCR1––CIDCR7" - line 820: An URL would be useful. - line 828: There are two dashes in between the page numbers. - This may be an issue with the reviewer version of the PDF, but clickable links, both internal (footnotes, citations), as well as external (dois/URLs) should work. - p.2, footnote 6: URL uses different font than other URLs (and probably should be an ssl-URL too) - p.6, footnotes 7011: URLs use a different font than URLs in footnotes 3-5. Given the length and subsequent unreadable line-break in the URL of footnote 11, I suggest to try the (smaller) font of the URLs 3-5 here. - Aside from line 766, all references seem to be missing their dois. - All figures seem to have been provided in png format. However, being a raster-format, this might produce low-quality results for various zoom-levels. Please consider using scalable formats instead for plots like the ones in this publication.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A SURVEY OF THE STATE OF THE PRACTICE FOR RESEARCH SOFTWARE IN THE UNITED STATES Review round: 2 Reviewer: 1

Basic reporting: The edits to the basic reporting show attention to detail and respect for the reviewer comments. The highlighted concerns seem to have been addressed. However, some of the edits could be improved as follows: 1. Figure 1 is difficult to read. The font size is too small for the titles of the subplots. 2. Figures 1 and 7 show dots that are confusing. The text does not say what the dots represent. They could be data points for individual responses, but then they don't show the multiplicity when more than one respondent has the same answer. Also, there are cases (like NIH for Developers and DoE for Developers in Figure 7) where there are no "data points" below the mean. This makes me think that I am misinterpreting the dots. They should either be explained, or, if they don't convey any useful information, removed. 3. Although the response to reviewers stated that a Conclusions section was added, there still isn't a Conclusions section. The authors should add a Conclusions section. Experimental design: The edits to the experimental design address the reviewer comments. Validity of the findings: The edits to this section more accurately reflect the story that is told by the data, and the limitations of the data. Additional comments: no comment

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: A SURVEY OF THE STATE OF THE PRACTICE FOR RESEARCH SOFTWARE IN THE UNITED STATES Review round: 3 Reviewer: 1

Basic reporting: My concerns have all been addressed. Experimental design: My concerns have all been addressed. Validity of the findings: My concerns have all been addressed. Additional comments: My concerns have all been addressed.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MULTI-MODAL AFFINE FUSION NETWORK FOR SOCIAL MEDIA RUMOR DETECTION Review round: 1 Reviewer: 1

Basic reporting: Summary - In this study, the authors proposed the multimodal affine fusion network (MAFN) combined with entity recognition, a new end-to-end framework that fuses multimodal features to detect rumors effectively. The MAFN mainly consists of four parts: the entity recognition enhanced textual feature extractor, the visual feature extractor, the multimodal affine fuser, and the rumor detector. The entity recognition enhanced textual feature extractor is responsible for extracting textual features that enhance semantics with entity recognition from posts. The visual feature extractor extracts visual features. The multimodal affine fuser extracts the three types of modal features and fuses them by the affine method, and it cooperates with the rumor detector to learn the representations for rumor detection to produce reliable fusion detection. Extensive experiments were conducted on the MAFN based on real Weibo and Twitter multimodal datasets, which verified the effectiveness of the proposed multimodal fusion neural network in rumor detection. 2. Strength - this paper firstly introduced the four modules of the proposed MAFN model, i.e., the entity recognition enhanced textual feature extractor, the visual feature extractor, the multi-modal affine fuser, and the rumor classifier. Secondly, this paper describes how to integrate the four modules to represent and detect rumors. 3. Weakness - Lack of novelty of research. textual feature extractor -based problem solving is a very common approach in the recent deep learning field, and post-processing is also difficult to consider as a new algorithm. - The part about the learning scenario is confusing. A more understandable explanation is needed for training and testing. - The entity resolution feature is quite confusing, and difficult to understand. 4. Minor comments - There is an error in the reference. I haven't looked at all of them in detail. - The manuscript is not well organized. The introduction section must introduce the status and motivation of this work and summarize with a paragraph about this paper. - What are the limitations of the related works -Are there any limitations of this carried out study? -How to select and optimize the user-defined parameters in the proposed model? Experimental design: please see above Validity of the findings: please see above Additional comments: please see above

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MULTI-MODAL AFFINE FUSION NETWORK FOR SOCIAL MEDIA RUMOR DETECTION Review round: 1 Reviewer: 2

Basic reporting: The authors proposed the multimodal affine fusion network (MAFN) combined with entity recognition, a new end-to-end framework that fuses multimodal features to detect rumors effectively. Experimental design: The MAFN mainly consists of four parts: the entity recognition enhanced textual feature extractor, the visual feature extractor, the multimodal affine fuser, and the rumor detector. The entity recognition enhanced textual feature extractor is responsible for extracting textual features that enhance semantics with entity recognition from posts. The visual feature extractor extracts visual features. The multimodal affine fuser extracts the three types of modal features and fuses them by the affine method, and it cooperates with the rumor detector to learn the representations for rumor detection to produce reliable fusion detection. Validity of the findings: The manuscript sounds technically good, I have the following concerns that should be addressed before any decision. Grammatical mistakes 1. However, without supervision, the authenticity of published information cannot be detected--> should be ... "However, the authenticity of published information cannot be detected without supervision". 2. With the rapid development of the Internet, people obtain abundant information from social media such as Twitter and Weibo every day. However, due to the complex structure of social media, many rumors with corresponding images are mixed in real information to be widely spread, which misleads readers and exerts negative effects on society --> should be---> "With the rapid development of the Internet, people obtain much information from social media such as Twitter and Weibo every day. However, due to the complex structure of social media, many rumors with corresponding images are mixed in real information to be widely spread, which misleads readers and exerts adverse effects on society." 3. Established based on the multi-modal affine fuser, the rumor detector sent the finally obtained multi-modal feature--> "Based on the multi-modal affine fuser, the rumor detector sent the finally obtained multi-modal feature." Minor Changes: 1. The author should provide only relevant information related to this paper and reserve more space for the proposed framework. 2. The theoretical perceptive of all the models used for comparison must be included in the literature. 3. What are the real-life use cases of the proposed model? The authors can add a theoretical discussion on the real-life usage of the proposed model. 4. However, the author should compare the proposed algorithm with other recent works or provide a discussion. Otherwise, it's hard for the reader to identify the novelty and contribution of this work. 5. The descriptions given in this proposed scheme are not sufficient that this manuscript only adopted a variety of existing methods to complete the experiment where there are no strong hypotheses and methodical theoretical arguments. Therefore, the reviewer considers that this paper needs more works. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MULTI-MODAL AFFINE FUSION NETWORK FOR SOCIAL MEDIA RUMOR DETECTION Review round: 1 Reviewer: 3

Basic reporting: The abstract should be reformulated. The abstract is an extremely important and powerful representation of the article. The authors have to clarify what is the novelty of this paper in abstract. • Reduce challenges list as much as you can. • Provide the related works clearly highlight the main gap. • Authors have proposed three algorithms, but I do not understand which one is used in the comparison with state-of-the-art works. • Figures 5abcd, have same label which quite confusing what is the point from each figure. • Figures looks very fuzzy, and resolution of image is poor. • Replace Case Study Performance Visualization with the discussion section as it is very poor and more deep discussion is needed for findings of the study. • Conclusion needs to be improved. The most important obtained results should be briefly and clearly mentioned through the support of numerical data in the conclusion. • The details in this manuscript are vague, especially in the depth feature extraction, which is the major defect of this paper. In addition, there is a gap between the experimental condition and the real scene. So this method can not be applied to the real scene effectively. • There are many words and figures about the background/architectures of the proposed networks used in this paper that can be omitted. These proposed networks are in the field for a while and they are known by most likely every researcher in the field. • There is a lack of comparison with other studies in the discussion. I do know that from the “related work” introduced in this paper that most previous study provides a very high accuracy/statistic in a much smaller dataset. The quantitative results are lower if just compare to the numeric values. However, the model in this study could be more robust than other previously published models applying your dataset using other models. Experimental design: see above Validity of the findings: see above Additional comments: see above

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MULTI-MODAL AFFINE FUSION NETWORK FOR SOCIAL MEDIA RUMOR DETECTION Review round: 2 Reviewer: 1

Basic reporting: I'm satisfied with the current version. Experimental design: I'm satisfied with the current version. Validity of the findings: I'm satisfied with the current version. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MULTI-MODAL AFFINE FUSION NETWORK FOR SOCIAL MEDIA RUMOR DETECTION Review round: 2 Reviewer: 2

Basic reporting: The paper seems improved as compared to previous versions. Hence, it is acceptable for publication. Experimental design: The paper seems improved as compared to previous versions. Hence, it is acceptable for publication. Validity of the findings: The paper seems improved as compared to previous versions. Hence, it is acceptable for publication. Additional comments: No additional comments

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: MULTI-MODAL AFFINE FUSION NETWORK FOR SOCIAL MEDIA RUMOR DETECTION Review round: 2 Reviewer: 3

Basic reporting: Thanks, although the manuscript is improved, however, there are some minor things that should be considered before publication. - For example, the organization of the paper. All tables and figures should be cited in order from low to high. Experimental design: Looks good to me Validity of the findings: Looks good to me Additional comments: - Some grammatical and punctuation issues exist in the paper. It should be rectified. - The figure quality still needs enhacments.

You are one of the reviewers, your task is to write a review for the article. You will be given the title of the article, the number of the round in which the article is located, and your order among the reviewers.

Title: DEEP LEARNING METHODS FOR INVERSE PROBLEMS Review round: 1 Reviewer: 1

Basic reporting: I suggest you clarify the innovation of this paper in the introduction, that is, the contribution made by this article. Secondly, this article contains a large number of formulas. I hope you can explain the meaning of each symbol in the formula in detail, so as not to cause trouble to readers. In addition, the figures in this paper need to be explained in more detail or clearer images are selected, as shown in FIG. 11 Formal results should include clear definitions of all terms and theorems, and detailed proofs. Experimental design: This article explores deep learning methods for inverse problems. In order to facilitate readers to implement the methods used in this article, I suggest that you need to clearly list the hyperparameters required for deep learning training in the article. The data set used in this study are available publically which is completely limited and imbalance. In this article, you use various quantitative indicators to measure the advantages and disadvantages of various methods. I hope you can describe in detail the principle of these indicators and their relationship with the quality of the model. For the overall beauty of the article, I suggest you use a unified style for the drawings of this article. Validity of the findings: The methods used in this article are the results of early work. I suggest you use the latest achievements to enrich the work of this article, such as the work of the last year. Additional comments: No comment. See the comments in the above three parts for details