English

Combining Textual and Structural Information for Premise Selection in Lean

Machine Learning 2025-12-02 v2 Artificial Intelligence Computation and Language Logic in Computer Science

Abstract

Premise selection is a key bottleneck for scaling theorem proving in large formal libraries. Yet existing language-based methods often treat premises in isolation, ignoring the web of dependencies that connects them. We present a graph-augmented approach that combines dense text embeddings of Lean formalizations with graph neural networks over a heterogeneous dependency graph capturing both state-premise and premise-premise relations. On the LeanDojo Benchmark, our method outperforms the ReProver language-based baseline by over 25\% across standard retrieval metrics. These results suggest that relational information is beneficial for premise selection.

Keywords

Cite

@article{arxiv.2510.23637,
  title  = {Combining Textual and Structural Information for Premise Selection in Lean},
  author = {Job Petrovčič and David Eliecer Narvaez Denis and Ljupčo Todorovski},
  journal= {arXiv preprint arXiv:2510.23637},
  year   = {2025}
}
R2 v1 2026-07-01T07:08:11.333Z