AlphaVerus: Bootstrapping Formally Verified Code Generation through Self-Improving Translation and Treefinement
Track: Theme Track: Agents & Reasoning
Keywords: formal verification, code generation, self-improvement, search+llms
TL;DR: AlphaVerus leverages self-improvement and formal verification to generate provably correct code by using exploration, code refinement via tree search, preventing reward hacking and demonstrating state-of-the-art verified code generation capabilities.
Abstract: Automated code generation with large language models has gained significant traction, but there remains no guarantee of the correctness of generated code. We aim to use formal verification to provide mathematical guarantees that the generated code is correct. However, generating formally verified code with LLMs is hindered by the scarcity of training data and the complexity of formal proofs. To tackle this challenge, we introduce AlphaVerus, a self-improving framework that bootstraps formally verified code generation by iteratively translating programs from a higher-resource language and leveraging feedback from a verifier. AlphaVerus operates in three phases: exploration of candidate translations, Treefinement -- a novel tree search algorithm for program refinement using verifier feedback, and filtering misaligned specifications and programs to prevent reward hacking. Through this iterative process, AlphaVerus enables LLaMA-3.1-70B model to generate verified code without human intervention or model finetuning. AlphaVerus shows an ability to generate formally verified solutions for HumanEval and MBPP, laying the groundwork for truly trustworthy code-generation agents.
Submission Number: 19
Loading