ORSMT: Unleashing the Potential of Satisfiability Modulo Theories for Semantic Verification and Repair in LLM-Based Optimization

Key points

• Verifying one optimization problem across these artifacts requires semantic comparison: unequal models can attain the same objective, equivalent models can use different auxiliary variables, and mutually consistent artifacts can omit the same source requirement.1
• ORSMT provides a typed semantic interface and exact SMT relation queries.1
• It aligns observable decisions, projects auxiliary variables, proves or refutes feasible region and objective relations, localizes refutations to implicated rows or domains and links source texts when available.1
• Across 3,034 problems and six generator sizes, ORSMT improves the model-code match rate in all 72 benchmark and generator combinations before selection.1

Sources (1)

• [1]ORSMT: Unleashing the Potential of Satisfiability Modulo Theories for Semantic Verification and Repair in LLM-Based Optimization
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 21, 03:27 AM

  Large language models increasingly automate optimization modeling from natural language problems, producing written mathematics, solver code and a returned report.

  Verifying one optimization problem across these artifacts requires semantic comparison: unequal models can attain the same objective, equivalent models can use different auxiliary variables, and mutually consistent artifacts can omit the same source requirement.

Extractive summary: sentences quoted from the sources.