Learning-Infused Formal Reasoning: From Contract Synthesis to Artifact Reuse and Formal Semantics
Arshad Beg, Diarmuid O’Donoghue, Rosemary Monahan · 2026
Read Learning-Infused Formal Reasoning: From Contract Synthesis to Artifact Reuse and Formal Semantics by Arshad Beg, Diarmuid O’Donoghue, Rosemary Monahan (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-2602-02881.txt.
Description
A long-term vision for proving software correct is laid out by Arshad Beg and 2 co-authors in a 2026 paper. Formal methods check that systems match their specifications. Today each proof is mostly built alone, then set aside. The authors argue that future tools should reuse past work instead.
Their framework pairs large language models with graph-based representations of specifications and proofs. Learning-based parts guide matching across different notations and levels of detail. Symbolic checks keep the results formally sound. Contracts could then be generated, transferred, and refined across systems over time. The paper reports early work toward that goal and maps the research still needed.
Catalog record
| CREATOR | Arshad Beg, Diarmuid O’Donoghue, Rosemary Monahan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-02881