NL-Debugging: Exploiting Natural Language as an Intermediate Representation for Code Debugging

2025-05-21

AI Summary

This paper addresses the challenge of diagnosing and repairing algorithmic logic errors in code debugging. We propose NL-Debugging, the first framework to employ natural language (NL) as a semantic-level intermediate representation for debugging. Our method integrates large language model–driven NL reasoning, execution-feedback-guided iterative NL refinement, and bidirectional code–NL mapping with consistency verification. Key contributions are: (1) establishing structured NL descriptions as the optimal intermediate representation for debugging; (2) empirically demonstrating NL reasoning’s substantial gains in localizing and fixing complex defects—including logic errors and missing boundary conditions; and (3) extending the modifiable search space beyond syntactic repair to ensure semantic consistency. Experiments show that NL-Debugging achieves a 32.7% higher repair success rate and reduces average debugging steps by 41% compared to state-of-the-art baselines.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Debugging is a critical aspect of LLM's coding ability. Early debugging efforts primarily focused on code-level analysis, which often falls short when addressing complex programming errors that require a deeper understanding of algorithmic logic. Recent advancements in large language models (LLMs) have shifted attention toward leveraging natural language reasoning to enhance code-related tasks. However, two fundamental questions remain unanswered: What type of natural language format is most effective for debugging tasks? And what specific benefits does natural language reasoning bring to the debugging process? In this paper, we introduce NL-DEBUGGING, a novel framework that employs natural language as an intermediate representation to improve code debugging. By debugging at a natural language level, we demonstrate that NL-DEBUGGING outperforms traditional debugging methods and enables a broader modification space through direct refinement guided by execution feedback. Our findings highlight the potential of natural language reasoning to advance automated code debugging and address complex programming challenges.

Problem

• Identify effective natural language formats for debugging
• Explore benefits of natural language reasoning in debugging
• Improve code debugging via natural language intermediate representation

Innovation

• Uses natural language as intermediate debugging representation
• Enables broader modification space via execution feedback
• Outperforms traditional code-level debugging methods