Reproduce “Towards Reliable C-to-Rust Translation with Rule-Guided Reasoning and Reinforcement Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Feng Luo, Jiachen Liu, Cuiyun Gao, Jia Feng, Kui Liu
TRAVEL translates C to Rust through rule-guided search, execution feedback, and a learned reasoning reward. It generates test inputs, explores translation paths, and uses reinforcement learning to improve its policy. On xCodeEval, Qwen2.5-Coder-7B-Instruct with TRAVEL reaches 71.07% computational accuracy and 87.38% compilation success, outperforming IRENE on both; its compilation rate remains below SFT, and tests do not prove equivalence.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.