Semantic-aware and Self-improving Program Reduction via Agentic Large Language Models
For developers debugging language processors, this work offers a more effective program reduction method that leverages LLM reasoning and self-improvement, though it is incremental in applying LLMs to a known task.
The paper tackles the problem of reducing bug-triggering programs to minimal forms for debugging language processors. The proposed method, PROJ, uses agentic LLMs to perform semantic-aware reduction and self-improvement, consistently producing smaller reduced programs than existing state-of-the-art reducers across 90 benchmarks in three programming languages.
Reducing bug-triggering programs to their minimal essential form is a fundamental task in debugging language processors such as compilers and interpreters. Existing reduction techniques are limited by their reliance on predefined, syntax-driven transformations that lack semantic understanding of the target program, and by their inability to learn from past reduction experiences. We present a new approach that recasts program reduction as an autonomous reasoning task powered by agentic Large Language Models (LLMs). Instead of applying fixed transformation rules, our method enables an LLM to analyze program semantics, formulate reduction hypotheses, and iteratively refine its approach based on execution outcomes. Successful reduction experiences are further distilled into reusable strategies, allowing the system to continuously improve over time. We realize this approach in PROJ, a framework built around two collaborative components: a reducer agent that performs semantic-aware, case-specific program reduction, and a reflector agent that extracts and accumulates transferable reduction knowledge. Extensive experiments on 90 benchmarks spanning three programming languages show that PROJ consistently produces smaller reduced programs than all existing state-of-the-art reducers while maintaining high efficiency.