Overview

Autoformalization translates informal mathematical statements, proofs, and software or scientific requirements into forms that proof assistants and solvers can check. Systems commonly combine language models with library retrieval, dependency-aware decomposition, and iterative feedback from Lean or Isabelle checkers, provers, model checkers, and tests. Benchmarks span competition and textbook mathematics, first-order logic, and executable software specifications, while agentic workflows increasingly tackle larger theories and research developments. The central challenge is semantic faithfulness: a formal artifact can compile or verify while misstating its source, so evaluation increasingly combines execution with equivalence checks, adversarial tests, and human review.

Latest papers

CardsList

124 papers17 papers in the last four weeks, up 21% on the four weeks before. 0.1% of all new papers.

  1. A Neurosymbolic Approach to Natural Language Formalization and Verification

    Nov 12, 2025Chenyang An, Sam Bayless, Stefano Buliani +27Formal VerificationLLM Safety

  2. TopoAlign: A Framework for Aligning Code to Math via Topological Decomposition

    Oct 13, 2025Yupei Li, Philipp Borchert, Gerasimos LampourasAutomated Theorem ProvingAutoformalization

  3. Aria: An Agent For Retrieval and Iterative Auto-Formalization via Dependency Graph

    Oct 6, 2025Hanyu Wang, Ruohan Xie, Yutong Wang +3Lean Theorem ProvingAutoformalization

  4. Beyond Gold Standards: Epistemic Ensemble of LLM Judges for Formal Mathematical Reasoning

    Jun 12, 2025Lan Zhang, Marco Valentino, Jordan Meadows +1Automated EvaluationAutoformalization