Update README.md
Browse files
README.md
CHANGED
@@ -8,7 +8,7 @@ license:
|
|
8 |
multilinguality:
|
9 |
- monolingual
|
10 |
size_categories:
|
11 |
-
- n<
|
12 |
source_datasets:
|
13 |
- extended|other
|
14 |
task_categories:
|
@@ -86,21 +86,41 @@ The data fields are the same among all splits.
|
|
86 |
- `token_idx_2`: a `int` feature.
|
87 |
- `date_2`: a `string` feature.
|
88 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
89 |
|
90 |
### Data Splits
|
91 |
|
92 |
| task | description | number of instances |
|
93 |
|:-----------------|:-----------------------------------|:----------------------|
|
94 |
-
|
|
95 |
| tweet_ner7 | sequence labeling | 4616 / 576 / 2807 |
|
96 |
| tweet_qa | generation | 9489 / 1086 / 1203 |
|
|
|
97 |
| tweet_similarity | regression on two texts | 450 / 100 / 450 |
|
98 |
-
| tweet_topic | multi-label classification | 4585 / 573 / 1679 |
|
99 |
| tempo_wic | binary classification on two texts | 1427 / 395 / 1472 |
|
100 |
-
| tweet_sentiment | ABSA on a five-pointscale | 26632 / 4000 / 12379 |
|
101 |
| tweet_hate | multi-class classification | 5019 / 716 / 1433 |
|
102 |
| tweet_emoji | multi-class classification | 50,000 / 5,000 / 50,000 |
|
103 |
-
|
|
|
|
104 |
|
105 |
## Citation Information
|
106 |
|
@@ -193,4 +213,57 @@ TBA
|
|
193 |
pages = "3353--3359",
|
194 |
abstract = "Language evolves over time, and word meaning changes accordingly. This is especially true in social media, since its dynamic nature leads to faster semantic shifts, making it challenging for NLP models to deal with new content and trends. However, the number of datasets and models that specifically address the dynamic nature of these social platforms is scarce. To bridge this gap, we present TempoWiC, a new benchmark especially aimed at accelerating research in social media-based meaning shift. Our results show that TempoWiC is a challenging benchmark, even for recently-released language models specialized in social media.",
|
195 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
196 |
```
|
|
|
8 |
multilinguality:
|
9 |
- monolingual
|
10 |
size_categories:
|
11 |
+
- n<50K
|
12 |
source_datasets:
|
13 |
- extended|other
|
14 |
task_categories:
|
|
|
86 |
- `token_idx_2`: a `int` feature.
|
87 |
- `date_2`: a `string` feature.
|
88 |
|
89 |
+
#### tweet_hate
|
90 |
+
- `gold_label`: a `int` feature.
|
91 |
+
- `text`: a `string` feature.
|
92 |
+
|
93 |
+
#### tweet_emoji
|
94 |
+
- `gold_label`: a `int` feature.
|
95 |
+
- `text`: a `string` feature.
|
96 |
+
|
97 |
+
#### tweet_sentiment
|
98 |
+
- `gold_label`: a `int` feature.
|
99 |
+
- `text`: a `string` feature.
|
100 |
+
- `target`: a `string` feature.
|
101 |
+
|
102 |
+
#### tweet_nerd
|
103 |
+
- `gold_label_binary`: a `int` feature.
|
104 |
+
- `target`: a `string` feature.
|
105 |
+
- `context`: a `string` feature.
|
106 |
+
- `definition`: a `string` feature.
|
107 |
+
- `text_start`: a `int` feature.
|
108 |
+
- `text_end`: a `int` feature.
|
109 |
|
110 |
### Data Splits
|
111 |
|
112 |
| task | description | number of instances |
|
113 |
|:-----------------|:-----------------------------------|:----------------------|
|
114 |
+
| tweet_topic | multi-label classification | 4585 / 573 / 1679 |
|
115 |
| tweet_ner7 | sequence labeling | 4616 / 576 / 2807 |
|
116 |
| tweet_qa | generation | 9489 / 1086 / 1203 |
|
117 |
+
| tweet_intimacy | regression on a single text | 1191 / 396 / 396 |
|
118 |
| tweet_similarity | regression on two texts | 450 / 100 / 450 |
|
|
|
119 |
| tempo_wic | binary classification on two texts | 1427 / 395 / 1472 |
|
|
|
120 |
| tweet_hate | multi-class classification | 5019 / 716 / 1433 |
|
121 |
| tweet_emoji | multi-class classification | 50,000 / 5,000 / 50,000 |
|
122 |
+
| tweet_sentiment | ABSA on a five-pointscale | 26632 / 4000 / 12379 |
|
123 |
+
| tweet_nerd | binary classification | * / 407 / * |
|
124 |
|
125 |
## Citation Information
|
126 |
|
|
|
213 |
pages = "3353--3359",
|
214 |
abstract = "Language evolves over time, and word meaning changes accordingly. This is especially true in social media, since its dynamic nature leads to faster semantic shifts, making it challenging for NLP models to deal with new content and trends. However, the number of datasets and models that specifically address the dynamic nature of these social platforms is scarce. To bridge this gap, we present TempoWiC, a new benchmark especially aimed at accelerating research in social media-based meaning shift. Our results show that TempoWiC is a challenging benchmark, even for recently-released language models specialized in social media.",
|
215 |
}
|
216 |
+
```
|
217 |
+
|
218 |
+
- TweetHate
|
219 |
+
```
|
220 |
+
@inproceedings{sachdeva-etal-2022-measuring,
|
221 |
+
title = "The Measuring Hate Speech Corpus: Leveraging Rasch Measurement Theory for Data Perspectivism",
|
222 |
+
author = "Sachdeva, Pratik and
|
223 |
+
Barreto, Renata and
|
224 |
+
Bacon, Geoff and
|
225 |
+
Sahn, Alexander and
|
226 |
+
von Vacano, Claudia and
|
227 |
+
Kennedy, Chris",
|
228 |
+
booktitle = "Proceedings of the 1st Workshop on Perspectivist Approaches to NLP @LREC2022",
|
229 |
+
month = jun,
|
230 |
+
year = "2022",
|
231 |
+
address = "Marseille, France",
|
232 |
+
publisher = "European Language Resources Association",
|
233 |
+
url = "https://aclanthology.org/2022.nlperspectives-1.11",
|
234 |
+
pages = "83--94",
|
235 |
+
abstract = "We introduce the Measuring Hate Speech corpus, a dataset created to measure hate speech while adjusting for annotators{'} perspectives. It consists of 50,070 social media comments spanning YouTube, Reddit, and Twitter, labeled by 11,143 annotators recruited from Amazon Mechanical Turk. Each observation includes 10 ordinal labels: sentiment, disrespect, insult, attacking/defending, humiliation, inferior/superior status, dehumanization, violence, genocide, and a 3-valued hate speech benchmark label. The labels are aggregated using faceted Rasch measurement theory (RMT) into a continuous score that measures each comment{'}s location on a hate speech spectrum. The annotation experimental design assigned comments to multiple annotators in order to yield a linked network, allowing annotator disagreement (perspective) to be statistically summarized. Annotators{'} labeling strictness was estimated during the RMT scaling, projecting their perspective onto a linear measure that was adjusted for the hate speech score. Models that incorporate this annotator perspective parameter as an auxiliary input can generate label- and score-level predictions conditional on annotator perspective. The corpus includes the identity group targets of each comment (8 groups, 42 subgroups) and annotator demographics (6 groups, 40 subgroups), facilitating analyses of interactions between annotator- and comment-level identities, i.e. identity-related annotator perspective.",
|
236 |
+
}
|
237 |
+
```
|
238 |
+
|
239 |
+
- TweetEmoji
|
240 |
+
```TBA``
|
241 |
+
|
242 |
+
- TweetSentiment
|
243 |
+
```
|
244 |
+
@inproceedings{rosenthal-etal-2017-semeval,
|
245 |
+
title = "{S}em{E}val-2017 Task 4: Sentiment Analysis in {T}witter",
|
246 |
+
author = "Rosenthal, Sara and
|
247 |
+
Farra, Noura and
|
248 |
+
Nakov, Preslav",
|
249 |
+
booktitle = "Proceedings of the 11th International Workshop on Semantic Evaluation ({S}em{E}val-2017)",
|
250 |
+
month = aug,
|
251 |
+
year = "2017",
|
252 |
+
address = "Vancouver, Canada",
|
253 |
+
publisher = "Association for Computational Linguistics",
|
254 |
+
url = "https://aclanthology.org/S17-2088",
|
255 |
+
doi = "10.18653/v1/S17-2088",
|
256 |
+
pages = "502--518",
|
257 |
+
abstract = "This paper describes the fifth year of the Sentiment Analysis in Twitter task. SemEval-2017 Task 4 continues with a rerun of the subtasks of SemEval-2016 Task 4, which include identifying the overall sentiment of the tweet, sentiment towards a topic with classification on a two-point and on a five-point ordinal scale, and quantification of the distribution of sentiment towards a topic across a number of tweets: again on a two-point and on a five-point ordinal scale. Compared to 2016, we made two changes: (i) we introduced a new language, Arabic, for all subtasks, and (ii) we made available information from the profiles of the Twitter users who posted the target tweets. The task continues to be very popular, with a total of 48 teams participating this year.",
|
258 |
+
}
|
259 |
+
```
|
260 |
+
|
261 |
+
- TweetNERD
|
262 |
+
```
|
263 |
+
@article{mishra2022tweetnerd,
|
264 |
+
title={TweetNERD--End to End Entity Linking Benchmark for Tweets},
|
265 |
+
author={Mishra, Shubhanshu and Saini, Aman and Makki, Raheleh and Mehta, Sneha and Haghighi, Aria and Mollahosseini, Ali},
|
266 |
+
journal={arXiv preprint arXiv:2210.08129},
|
267 |
+
year={2022}
|
268 |
+
}
|
269 |
```
|