https://github.com/declare-lab/cicero
The purpose of this repository is to introduce new dialogue-level commonsense inference datasets and tasks. We chose dialogues as the data source because dialogues are known to be complex and rich in commonsense.
https://github.com/declare-lab/cicero
commonsense-reasoning dialogue-systems roberta-model seq2seq transformers
Last synced: 10 months ago
JSON representation
The purpose of this repository is to introduce new dialogue-level commonsense inference datasets and tasks. We chose dialogues as the data source because dialogues are known to be complex and rich in commonsense.
- Host: GitHub
- URL: https://github.com/declare-lab/cicero
- Owner: declare-lab
- License: mit
- Created: 2022-02-26T02:08:02.000Z (almost 4 years ago)
- Default Branch: main
- Last Pushed: 2023-03-14T10:08:25.000Z (almost 3 years ago)
- Last Synced: 2025-03-27T18:21:42.331Z (11 months ago)
- Topics: commonsense-reasoning, dialogue-systems, roberta-model, seq2seq, transformers
- Language: Python
- Homepage:
- Size: 7.62 MB
- Stars: 63
- Watchers: 4
- Forks: 6
- Open Issues: 8
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
:fire::fire: We have released the checkpoint of our dialogue-level transformer model DIALeCT for contextual commonsense inference generation on Huggingface along with a demo. Check it out: https://huggingface.co/declare-lab/dialect
# Contextualized Commonsense Inference in Dialogues (CICERO)
The purpose of this repository is to introduce new dialogue-level commonsense inference datasets and tasks. We chose dialogues as the data source because dialogues are known to be complex and rich in commonsense. At present, we have released two versions of the dataset:
## CICERO-v1
CICERO-v1 can be found here: https://github.com/declare-lab/CICERO/tree/main/v1.
In this dataset, each training instance is associated with only one human-written inference. There are two tasks pertaining to this dataset: 1) generative commonsense inference in dialogues, and 2) multiple choice answer selection.
## CICERO-v2
Depending on a situation, multiple different reasonings are possible each leading to various unique inferences. In constructing CICERO-v2, we asked annotators to write more than one plausible inference for each dialogue context. We call this task --- Multiview Contextual Commonsense Inference, a highly challenging task for large language models. CICERO-v2 is available here: https://github.com/declare-lab/CICERO/tree/main/v2.
## Citation
If these datasets are useful in your research, please cite the following papers:
### CICERO-v1
```
@inproceedings{ghosal-etal-2022-cicero,
title = "{CICERO}: A Dataset for Contextualized Commonsense Inference in Dialogues",
author = "Ghosal, Deepanway and
Shen, Siqi and
Majumder, Navonil and
Mihalcea, Rada and
Poria, Soujanya",
booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = may,
year = "2022",
address = "Dublin, Ireland",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.acl-long.344",
pages = "5010--5028",
}
```
### CICERO-v2
```
@inproceedings{shen-et-al-cicerov2-2022,
title = "Multiview Contextual Commonsense Inference: A New Dataset and Task",
author = "Shen, Siqi and
Ghosal, Deepanway and
Majumder, Navonil and
Lim, Henry and
Mihalcea, Rada and
Poria, Soujanya",
journal = "arxiv"
}
```