Legal2LogicICL: Improving Generalization in Transforming Legal Cases to Logical Formulas via Diverse Few-Shot Learning Conference: 21th International Conference on Artificial Intelligence and Law; June 8–16, 2026; Singapore CCS: Software and its engineering Semantics CCS: Applied computing Law
Jieying Xue, Phuong Minh Nguyen, Ha Thanh Nguyen, May Myo Zin, Ken Satoh · 2026
Read Legal2LogicICL: Improving Generalization in Transforming Legal Cases to Logical Formulas via Diverse Few-Shot Learning Conference: 21th International Conference on Artificial Intelligence and Law; June 8–16, 2026; Singapore CCS: Software and its engineering Semantics CCS: Applied computing Law by Jieying Xue, Phuong Minh Nguyen, Ha Thanh Nguyen, May Myo Zin, Ken Satoh (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-11699.txt.
Description
Xue and four colleagues improve reach in turning legal cases into logic formulas with diverse few-shot learning. Logic-based reasoning fine-tunes models to map cases into formulas, but annotated training data are scarce. Legal2LogicICL enables in-context learning via retrieval, balancing exemplar mix at semantic levels. It also cuts entity bias, where long entity mentions swamp semantics and obscure legal patterns. The framework builds few-shot demos for logical rule making without extra training, plus a Legal2Proleg dataset linking cases with PROLEG formulas.
Catalog record
| CREATOR | Jieying Xue, Phuong Minh Nguyen, Ha Thanh Nguyen, May Myo Zin, Ken Satoh |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54325 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.11699