--- language: - en library_name: transformers widget: - text: >- X logs in from a company computer xEffect 10 seconds later [GEN] license: cc-by-nc-4.0 --- # TaCOMET_en This is the English TaCOMET model, which is the finetuned COMET model on the English ver. of [TimeATOMIC](https://github.com/nlp-waseda/TaCOMET) using causal language modeling (CLM) objective. We use [COMET-distil](https://github.com/peterwestai2/symbolic-knowledge-distillation), introduced in this [paper](https://aclanthology.org/2022.naacl-main.341/), as the base COMET model. The data and this model are introduced in [LREC-COLING2024](https://aclanthology.org/2024.lrec-main.1405/). ### BibTeX entry and citation info ```bibtex @inproceedings{murata-kawahara-2024-time-aware, title = "Time-aware {COMET}: A Commonsense Knowledge Model with Temporal Knowledge", author = "Murata, Eiki and Kawahara, Daisuke", editor = "Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen", booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", month = may, year = "2024", address = "Torino, Italia", publisher = "ELRA and ICCL", url = "https://aclanthology.org/2024.lrec-main.1405", pages = "16162--16174", abstract = "To better handle commonsense knowledge, which is difficult to acquire in ordinary training of language models, commonsense knowledge graphs and commonsense knowledge models have been constructed. The former manually and symbolically represents commonsense, and the latter stores these graphs{'} knowledge in the models{'} parameters. However, the existing commonsense knowledge models that deal with events do not consider granularity or time axes. In this paper, we propose a time-aware commonsense knowledge model, TaCOMET. The construction of TaCOMET consists of two steps. First, we create TimeATOMIC using ChatGPT, which is a commonsense knowledge graph with time. Second, TaCOMET is built by continually finetuning an existing commonsense knowledge model on TimeATOMIC. TimeATOMIC and continual finetuning let the model make more time-aware generations with rich commonsense than the existing commonsense models. We also verify the applicability of TaCOMET on a robotic decision-making task. TaCOMET outperformed the existing commonsense knowledge model when proper times are input. Our dataset and models will be made publicly available.", } ```