Fully Autonomous Programming using Iterative Multi-Agent Debugging with Large Language Models

• URL: http://arxiv.org/abs/2503.07693v1
• Date: Mon, 10 Mar 2025 16:56:51 GMT
• Title: Fully Autonomous Programming using Iterative Multi-Agent Debugging with Large Language Models
• Authors: Anastasiia Grishina, Vadim Liventsev, Aki Härmä, Leon Moonen,
• Abstract summary: Program synthesis with Large Language Models (LLMs) suffers from a "near-miss syndrome"<n>We address this with a multi-agent framework called Synthesize, Execute, Instruct, Debug, and Repair (SEIDR)<n>We empirically explore these trade-offs by comparing replace-focused, repair-focused, and hybrid debug strategies.
• Score: 8.70160958177614
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Program synthesis with Large Language Models (LLMs) suffers from a "near-miss syndrome": the generated code closely resembles a correct solution but fails unit tests due to minor errors. We address this with a multi-agent framework called Synthesize, Execute, Instruct, Debug, and Repair (SEIDR). Effectively applying SEIDR to instruction-tuned LLMs requires determining (a) optimal prompts for LLMs, (b) what ranking algorithm selects the best programs in debugging rounds, and (c) balancing the repair of unsuccessful programs with the generation of new ones. We empirically explore these trade-offs by comparing replace-focused, repair-focused, and hybrid debug strategies. We also evaluate lexicase and tournament selection to rank candidates in each generation. On Program Synthesis Benchmark 2 (PSB2), our framework outperforms both conventional use of OpenAI Codex without a repair phase and traditional genetic programming approaches. SEIDR outperforms the use of an LLM alone, solving 18 problems in C++ and 20 in Python on PSB2 at least once across experiments. To assess generalizability, we employ GPT-3.5 and Llama 3 on the PSB2 and HumanEval-X benchmarks. Although SEIDR with these models does not surpass current state-of-the-art methods on the Python benchmarks, the results on HumanEval-C++ are promising. SEIDR with Llama 3-8B achieves an average pass@100 of 84.2%. Across all SEIDR runs, 163 of 164 problems are solved at least once with GPT-3.5 in HumanEval-C++, and 162 of 164 with the smaller Llama 3-8B. We conclude that SEIDR effectively overcomes the near-miss syndrome in program synthesis with LLMs.

Related papers

• Code Generation with Small Language Models: A Deep Evaluation on Codeforces [2.314213846671956]
  Small Language Models offer faster inference, lower deployment overhead, and better adaptability to domain-specific tasks. We benchmark five open SLMs across 280 Codeforces problems spanning Elo ratings from 800 to 2100. PHI-4 14B achieved the best performance among SLMs, with a pass@3 of 63.6%.
  arXiv  Detail & Related papers  (2025-04-09T23:57:44Z)
• CodeElo: Benchmarking Competition-level Code Generation of LLMs with Human-comparable Elo Ratings [70.95565672516979]
  Existing benchmarks, like LiveCodeBench and USACO, fall short due to the unavailability of private test cases, lack of support for special judges, and misaligned execution environments.<n>CodeElo is a standardized competition-level code generation benchmark that effectively addresses all these challenges for the first time.
  arXiv  Detail & Related papers  (2025-01-02T13:49:00Z)
• Not All Votes Count! Programs as Verifiers Improve Self-Consistency of Language Models for Math Reasoning [24.386388107656334]
  We introduce Prove, a framework that leverages translated programs derived from natural language solutions as a verification mechanism.<n>Unlike vanilla majority voting, our approach filters out solutions whose corresponding program output is inconsistent with the generated solution, aggregating only those that pass verification.<n>Our results show that Prove consistently outperforms vanilla majority voting for solving mathematical reasoning tasks across all model sizes and datasets.
  arXiv  Detail & Related papers  (2024-10-16T14:24:55Z)
• LLaMA-Berry: Pairwise Optimization for O1-like Olympiad-Level Mathematical Reasoning [56.273799410256075]
  The framework combines Monte Carlo Tree Search (MCTS) with iterative Self-Refine to optimize the reasoning path. The framework has been tested on general and advanced benchmarks, showing superior performance in terms of search efficiency and problem-solving capability.
  arXiv  Detail & Related papers  (2024-10-03T18:12:29Z)
• What's Wrong with Your Code Generated by Large Language Models? An Extensive Study [80.18342600996601]
  Large language models (LLMs) produce code that is shorter yet more complicated as compared to canonical solutions. We develop a taxonomy of bugs for incorrect codes that includes three categories and 12 sub-categories, and analyze the root cause for common bug types. We propose a novel training-free iterative method that introduces self-critique, enabling LLMs to critique and correct their generated code based on bug types and compiler feedback.
  arXiv  Detail & Related papers  (2024-07-08T17:27:17Z)
• Code Repair with LLMs gives an Exploration-Exploitation Tradeoff [16.80314690163063]
  Iteratively improving and repairing source code with large language models (LLMs) has emerged as a popular way of generating programs that would be too complex to construct in one shot. We show here that refinement exposes an explore-exploit tradeoff: exploit by refining the program that passes the most test cases, or explore by refining a lesser considered program.
  arXiv  Detail & Related papers  (2024-05-26T04:00:30Z)
• ALGO: Synthesizing Algorithmic Programs with LLM-Generated Oracle Verifiers [60.6418431624873]
  Large language models (LLMs) excel at implementing code from functionality descriptions but struggle with algorithmic problems. We propose ALGO, a framework that synthesizes Algorithmic programs with LLM-Generated Oracles to guide the generation and verify their correctness. Experiments show that when equipped with ALGO, we achieve an 8x better one-submission pass rate over the Codex model and a 2.6x better one-submission pass rate over CodeT.
  arXiv  Detail & Related papers  (2023-05-24T00:10:15Z)
• Fully Autonomous Programming with Large Language Models [0.9558392439655015]
  Current approaches to program synthesis with Large Language Models (LLMs) exhibit a "near miss syndrome" We use OpenAI Codex as the LLM and Program Synthesis Benchmark 2 as a database of problem descriptions and tests for evaluation. The resulting framework outperforms both conventional usage of Codex without the repair phase and traditional genetic programming approaches.
  arXiv  Detail & Related papers  (2023-04-20T16:12:05Z)
• PAL: Program-aided Language Models [112.94785609781503]
  We present Program-Aided Language models (PaL) to understand natural language problems. PaL offloads the solution step to a programmatic runtime such as a Python interpreter. We set new state-of-the-art results in all 12 benchmarks.
  arXiv  Detail & Related papers  (2022-11-18T18:56:13Z)
• Learning from Self-Sampled Correct and Partially-Correct Programs [96.66452896657991]
  We propose to let the model perform sampling during training and learn from both self-sampled fully-correct programs and partially-correct programs. We show that our use of self-sampled correct and partially-correct programs can benefit learning and help guide the sampling process. Our proposed method improves the pass@k performance by 3.1% to 12.3% compared to learning from a single reference program with MLE.
  arXiv  Detail & Related papers  (2022-05-28T03:31:07Z)

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.