Semantic-aware and Self-improving Program Reduction via Agentic Large Language Models

• URL: http://arxiv.org/abs/2607.03766v1
• Date: Sat, 04 Jul 2026 08:36:10 GMT
• Title: Semantic-aware and Self-improving Program Reduction via Agentic Large Language Models
• Abstract summary: We present a new approach that recasts program reduction as an autonomous reasoning task powered by agentic Large Language Models (LLMs)<n>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.
• Score: 5.1805580173652706
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• Beyond the Traceback: Using LLMs for Adaptive Explanations of Programming Errors [52.619793817715326]
  We present a multi-stage crowdsourced study evaluating skill-targeted, LLM-generated Python error messages.<n>We show that while LLM-rewritten messages significantly improved subjective evaluations, these perceived gains did not translate into statistically significant improvements in objective debug performance.<n>This highlights a critical human-AI complementarity gap: explanations that feel better to users do not necessarily make them more effective debuggers.
  arXiv  Detail & Related papers  (2026-08-21T09:14:28Z)
• Procedural Refinement by LLM-driven Algorithmic Debugging for ARC-AGI-2 [10.013415375696557]
  We propose a neuro-symbolic procedural refinement approach, Abduction-Based Procedural Refinement (ABPR)<n>ABPR couples an LLM with a meta-interpreter that materialises program execution into compact, declarative tree-structured traces.<n>Our experiments show that ABPR paired with Gemini-3-Flash a Pass@2 score of 56.67% even in a language in which contemporary LLMs typically underperform.
  arXiv  Detail & Related papers  (2026-03-20T08:16:15Z)
• Understanding by Reconstruction: Reversing the Software Development Process for LLM Pretraining [66.89012795621349]
  Large Language Models (LLMs) often struggle with the deep, long-horizon reasoning required for complex software engineering.<n>We propose a novel paradigm: understanding via reconstruction.<n>We introduce a framework that synthesizes latent agentic trajectories using a multi-agent simulation.
  arXiv  Detail & Related papers  (2026-03-11T09:23:20Z)
• Grammar-Guided Evolutionary Search for Discrete Prompt Optimisation [63.97051732013936]
  We propose an evolutionary search approach to automated discrete prompt optimisation consisting of two phases.<n>In the first phase, grammar-guided genetic programming is invoked to synthesise prompt-creating programmes.<n>In the second phase, local search is applied to explore the neighbourhoods of best-performing programmes.
  arXiv  Detail & Related papers  (2025-07-14T14:34:15Z)
• RePair: Automated Program Repair with Process-based Feedback [28.017321930042694]
  We show how small-scale language models (LM) can achieve excellent performance through process supervision and feedback. We develop a reward model that serves as a critic, providing feedback for the fine-tuned LM's action. The results show that process-based not only outperforms larger outcome-based generation methods, but also nearly matches the performance of closed-source commercial large-scale LMs.
  arXiv  Detail & Related papers  (2024-08-21T02:53:23Z)
• Peer-aided Repairer: Empowering Large Language Models to Repair Advanced Student Assignments [26.236420215606238]
  We develop a framework called PaR that is powered by the Large Language Model. PaR works in three phases: Peer Solution Selection, Multi-Source Prompt Generation, and Program Repair. The evaluation on Defects4DS and another well-investigated ITSP dataset reveals that PaR achieves a new state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-04-02T09:12:21Z)
• Moving beyond Deletions: Program Simplification via Diverse Program Transformations [11.038120567076772]
  Developers manually simplify program (known as developer-induced program simplification in this paper) to reduce its code size yet preserving its functionality. To reduce manual effort, rule-based approaches (e.g., deletion-based approaches) can be potentially applied to automate developer-induced program simplification. We propose SimpT5, a tool that can automatically produce simplified programs (semantically-equivalent programs with reduced source lines of code) Our evaluation shows that SimpT5 are more effective than prior approaches in automating developer-induced program simplification.
  arXiv  Detail & Related papers  (2024-01-26T22:59:43Z)
• CodeRL: Mastering Code Generation through Pretrained Models and Deep Reinforcement Learning [92.36705236706678]
  "CodeRL" is a new framework for program synthesis tasks through pretrained LMs and deep reinforcement learning. During inference, we introduce a new generation procedure with a critical sampling strategy. For the model backbones, we extended the encoder-decoder architecture of CodeT5 with enhanced learning objectives.
  arXiv  Detail & Related papers  (2022-07-05T02:42:15Z)
• Procedures as Programs: Hierarchical Control of Situated Agents through Natural Language [81.73820295186727]
  We propose a formalism of procedures as programs, a powerful yet intuitive method of representing hierarchical procedural knowledge for agent command and control. We instantiate this framework on the IQA and ALFRED datasets for NL instruction following.
  arXiv  Detail & Related papers  (2021-09-16T20:36:21Z)
• Learning from Executions for Semantic Parsing [86.94309120789396]
  We focus on the task of semi-supervised learning where a limited amount of annotated data is available. We propose to encourage executable programs for unlabeled utterances.
  arXiv  Detail & Related papers  (2021-04-12T21:07:53Z)
• Incremental maintenance of overgrounded logic programs with tailored simplifications [0.966840768820136]
  We introduce a new strategy for generating series of monotonically growing propositional programs. With respect to earlier approaches, our tailored simplification technique reduces the size of instantiated programs.
  arXiv  Detail & Related papers  (2020-08-06T21:50:11Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.