Beyond Code Pairs: Dialogue-Based Data Generation for LLM Code Translation

• URL: http://arxiv.org/abs/2512.03086v1
• Date: Sat, 29 Nov 2025 05:26:53 GMT
• Title: Beyond Code Pairs: Dialogue-Based Data Generation for LLM Code Translation
• Authors: Le Chen, Nuo Xu, Winson Chen, Bin Lei, Pei-Hung Lin, Dunzhi Zhou, Rajeev Thakur, Caiwen Ding, Ali Jannesari, Chunhua Liao,
• Abstract summary: We present an automated dataset generation pipeline featuring a dual-LLM Questioner-r design.<n>We show this data enables a 7B open-weight model to significantly outperform larger proprietary systems on key metrics like compilation success.
• Score: 22.50538010082899
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) have shown remarkable capabilities in code translation, yet their performance deteriorates in low-resource programming domains such as Fortran and emerging frameworks like CUDA, where high-quality parallel data are scarce. We present an automated dataset generation pipeline featuring a dual-LLM Questioner-Solver design that incorporates external knowledge from compilers and runtime feedback. Beyond traditional source-target code pair datasets, our approach additionally generates (1) verified translations with unit tests for assessing functional consistency, and (2) multi-turn dialogues that capture the reasoning process behind translation refinement. Applied to Fortran -> C++ and C++ -> CUDA, the pipeline yields 3.64k and 3.93k dialogues, respectively. Fine-tuning on this data yields dramatic improvements in functional correctness, boosting unit test success rates by over 56% on the challenging C++-to-CUDA task. We show this data enables a 7B open-weight model to significantly outperform larger proprietary systems on key metrics like compilation success.

Related papers

• Multi-Agent Systems for Dataset Adaptation in Software Engineering: Capabilities, Limitations, and Future Directions [8.97512410819274]
  This paper presents the first empirical study on how state-of-the-art multi-agent systems perform in dataset adaptation tasks.<n>We evaluate GitHub Copilot on adapting SE research artifacts from benchmark repositories including ROCODE and LogHub2.0.<n>Results show that current systems can identify key files and generate partial adaptations but rarely produce correct implementations.
  arXiv  Detail & Related papers  (2025-11-26T13:26:11Z)
• BRIDGE: Building Representations In Domain Guided Program Verification [67.36686119518441]
  BRIDGE decomposes verification into three interconnected domains: Code, Specifications, and Proofs.<n>We show that this approach substantially improves both accuracy and efficiency beyond standard error feedback methods.
  arXiv  Detail & Related papers  (2025-11-26T06:39:19Z)
• QiMeng-NeuComBack: Self-Evolving Translation from IR to Assembly Code [52.66657751895655]
  Large Language Models (LLMs) offer a compelling new paradigm: Neural Compilation.<n>This paper introduces NeuComBack, a novel benchmark dataset specifically designed for IR-to-assembly compilation.<n>We propose a self-evolving prompt optimization method that enables LLMs to evolve their internal prompt strategies.
  arXiv  Detail & Related papers  (2025-11-03T03:20:26Z)
• Loong: Synthesize Long Chain-of-Thoughts at Scale through Verifiers [103.4410890572479]
  We introduce the Loong Project: an open-source framework for scalable synthetic data generation and verification.<n>LoongBench is a curated seed dataset containing 8,729 human-vetted examples across 12 domains.<n>LoongEnv is a modular synthetic data generation environment that supports multiple prompting strategies to produce new question-answer-code triples.
  arXiv  Detail & Related papers  (2025-09-03T06:42:40Z)
• Function-to-Style Guidance of LLMs for Code Translation [59.487054943812836]
  We propose F2STrans, a function-to-style guiding paradigm designed to improve the performance of large language models in code translation.<n>Our approach comprises two key stages: (1) Functional learning, which optimize translation correctness using high-quality source-target code pairs.<n>We introduce a novel code translation benchmark that includes up-to-date source code, extensive test cases, and manually annotated ground-truth translations.
  arXiv  Detail & Related papers  (2025-07-15T08:25:02Z)
• OpenCodeReasoning-II: A Simple Test Time Scaling Approach via Self-Critique [59.18475981916166]
  We introduce OpenCodeReasoning-II, a dataset consisting of 2.5M question-solution-critique triples (approx. 35K unique programming questions)<n>In this work, we employ a two-stage supervised fine-tuning strategy. The first stage focuses on fine-tuning for code generation, while the second stage involves the joint training of models for both code generation and critique. Notably, the integration of our code generation and critique models leads to significant improvements in competitive coding performance.
  arXiv  Detail & Related papers  (2025-07-11T23:35:54Z)
• QiMeng-MuPa: Mutual-Supervised Learning for Sequential-to-Parallel Code Translation [27.60208311262231]
  Data scarcity poses a significant challenge for machine learning-based sequential-to-parallel code translation.<n>We propose textbfQiMeng-MuPa, a novel textbfMutual-Supervised Learning framework for Sequential-to-textbfParallel code translation.
  arXiv  Detail & Related papers  (2025-06-11T13:50:29Z)
• Thinking Longer, Not Larger: Enhancing Software Engineering Agents via Scaling Test-Time Compute [61.00662702026523]
  We propose a unified Test-Time Compute scaling framework that leverages increased inference-time instead of larger models.<n>Our framework incorporates two complementary strategies: internal TTC and external TTC.<n>We demonstrate our textbf32B model achieves a 46% issue resolution rate, surpassing significantly larger models such as DeepSeek R1 671B and OpenAI o1.
  arXiv  Detail & Related papers  (2025-03-31T07:31:32Z)
• Fortran2CPP: Automating Fortran-to-C++ Translation using LLMs via Multi-Turn Dialogue and Dual-Agent Integration [10.985254527043429]
  Our dataset comprises 11.7k dialogues capturing feedback-decision including code translation, compilation, execution, unit testing, and error-fixing.<n>Using this dataset, we achieve up to a 3.31x improvement in CodeBLEU scores and a 92% increase in compilation success rate.
  arXiv  Detail & Related papers  (2024-12-27T18:06:25Z)
• Enhancing Cross-Language Code Translation via Task-Specific Embedding Alignment in Retrieval-Augmented Generation [1.64043572114825]
  We introduce a novel method to enhance cross-language code translation from Fortran to C++ by integrating task-specific embedding alignment.<n>Our strategy aligns the retrieval model directly with the objective of maximizing translation quality, as quantified by the CodeBLEU metric.<n>By integrating these CodeBLEU-optimized embeddings into the RAG framework, our approach significantly enhances both retrieval accuracy and code generation quality.
  arXiv  Detail & Related papers  (2024-12-06T16:22:32Z)
• LLM-Assisted Code Cleaning For Training Accurate Code Generators [53.087019724256606]
  We investigate data quality for code and find that making the code more structured and readable leads to improved code generation performance of the system. We build a novel data-cleaning pipeline that uses these principles to transform existing programs. We evaluate our approach on two challenging algorithmic code generation benchmarks and find that fine-tuning CodeLLaMa-7B improves the performance by up to 30% compared to fine-tuning on the original dataset.
  arXiv  Detail & Related papers  (2023-11-25T02:45:50Z)
• Data Augmentation for Code Translation with Comparable Corpora and Multiple References [21.754147577489764]
  We build and analyze multiple types of comparable corpora, including programs generated from natural language documentation. To reduce overfitting to a single reference translation, we automatically generate additional translation references for available parallel data. Experiments show that our data augmentation techniques significantly improve CodeT5 for translation between Java, Python, and C++ by an average of 7.5% Computational Accuracy.
  arXiv  Detail & Related papers  (2023-11-01T06:01:22Z)

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.