nguyenthanhasia/SimpleLART

Dataset Description

SimpleLART is a quickly created dataset inspired by the paper How well do SOTA legal reasoning models support abductive reasoning?. It is designed to simplify the process of working with datasets without requiring complex libraries or multilingual setups. SimpleLART shares characteristics and the number of samples similar to the L'ART dataset described in the paper, with the following notes:

• Instead of generating samples randomly via theory generation using Lisp, this dataset utilizes samples from the latest released dataset by RuleTaker. If desired, users can also generate their own samples by following the RuleTaker guide: RuleTaker GitHub Repository.
• To our knowledge, the alpha NLI dataset's test set is no longer available (AlphaNLI on arXiv). However, the dataset reflects the state of the alpha NLI dataset at the time the paper was published.
• As we are unsure about the licenses of the original datasets, downloading this dataset implies that users must restrict its usage to research purposes only and respect the intellectual property (IP) rights of the original datasets.

Dataset Details

• Language(s) (NLP): English
• License: Research-only usage; respect original datasets' IP.

Dataset Sources

• Repository: RuleTaker GitHub Repository
• Paper: Alpha NLI on arXiv

Uses

Direct Use

SimpleLART is suitable for:

• Research in abductive reasoning.
• Testing and validating state-of-the-art (SOTA) legal reasoning models.
• Simplified dataset experiments without complex generation requirements.

Out-of-Scope Use

This dataset should not be used for commercial purposes or in any way that violates the intellectual property rights of the original datasets.

Dataset Structure

SimpleLART's structure mirrors the characteristics of the L'ART dataset, with samples derived or recreated as described above. Any deviations from the original datasets are noted in the Dataset Description.

Dataset Creation

Curation Rationale

SimpleLART was created to provide researchers with a lightweight, easy-to-access dataset that avoids the need for specialized tools or multiple programming languages for dataset generation.

Source Data

• Theory samples are sourced from the latest released dataset by RuleTaker.
• Samples reflecting alpha NLI data are included as they were available at the time the referenced paper was published.

Bias, Risks, and Limitations

• The dataset inherits biases, limitations, and risks of the original RuleTaker and AlphaNLI datasets.
• Usage is strictly limited to research and educational purposes.

Recommendations

Researchers should review the original datasets' licenses and adhere to the IP rights when using SimpleLART.

Citation

BibTeX:

@inproceedings{DBLP:conf/iclp/NguyenGTSS23,
  author={Ha-Thanh Nguyen and Randy Goebel and Francesca Toni and Kostas Stathis and Ken Satoh},
  title={How well do SOTA legal reasoning models support abductive reasoning?},
  year={2023},
  cdate={1672531200000},
  url={https://ceur-ws.org/Vol-3437/paper1LPLR.pdf},
  booktitle={ICLP Workshops},
  crossref={conf/iclp/2023w}
}