International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | ML-Insight: A Multi-Agent Collaborative Framework for Autonomous Debugging and Validation of Machine Learning Pipelines |
|---|---|
| ABSTRACT | Multi-agent AI system to autonomously inspect machine learning pipelines and models uploaded by users. The system analyzes execution logs, validates data schemas, checks feature compatibility, and evaluates model configurations to identify failures, inefficiencies, and potential risks. Specialized agents collaborate to recommend corrective actions, generate improved model variants, and automatically create preventive unit tests to reduce recurrence of similar issues. User interaction occurs through a web interface where ML artifacts are parsed and evaluated using Python scripts. Agent-based reasoning is leveraged instead of training new ML models to improve reliability, robustness, and performance of end-to-end ML pipelines. REST API integration enables seamless communication between agents and the interface, while automated testing continuously validates model improvements. The system streamlines ML pipeline debugging, enhances model quality, and supports developers in building reliable machine learning applications. |
| AUTHOR | R. SUBASHINI, DR. M. SENTHILKUMAR, GOKUL M, SELVA S, SHARVESH R Assistant Professor, Department of Artificial Intelligence and Data Science, Kongunadu College of Engineering and Technology, Trichy, India Professor, Department of Artificial Intelligence and Data Science, Kongunadu College of Engineering and Technology, Trichy, India Department of Artificial Intelligence and Data Science, Kongunadu College of Engineering and Technology, Trichy, India |
| VOLUME | 182 |
| DOI | DOI: 10.15680/IJIRCCE.2026.1403017 |
| pdf/17_ML-Insight A Multi-Agent Collaborative Framework for Autonomous Debugging and Validation of Machine Learning Pipelines.pdf | |
| KEYWORDS | |
| References | [1]. LangChain Blog, "Monte Carlo: Building Data + AI Observability Agents with LangGraph and LangSmith," Sep. 2025. [Online]. Available: https://blog.langchain.com/customers-monte-carlo/ [2]. IEEE Xplore, "Multi-Agent LLM Collaboration for Adaptive Code Review, Debugging, and Security Analysis," in Proc. IEEE International Conference on Software Engineering, 2025. doi: 10.1109/ICSE.2025.11135756 [3]. saim-x, "SAGE - Sequential Agent Goal Execution Protocol," GitHub Repository, Apr. 2025. [Online]. Available: https://github.com/saim-x/SAGE [4]. S. Rajan, "Multi-Agent Code Verification via Information Theory," arXiv preprint arXiv:2511.16708, Oct. 2025. [5]. IEEE Xplore, "Toward Generative 6G Simulation: An Experimental Multi-Agent LLM and ns-3 Integration," in Proc. IEEE International Conference on Communications, Jul. 2025. doi: 10.1109/ICC.2025.11104374 [6]. S. Jasper et al., "BugGen: A Self-Correcting Multi-Agent LLM Pipeline for Realistic RTL Bug Synthesis," arXiv preprint arXiv:2506.10501, Jun. 2025. [7]. S. Joshi, "LLMOps, AgentOps, and MLOps for Generative AI: A Comprehensive Review," PhilArchive, Jun. 2025. [8]. "FROAV: A Framework for RAG Observation and Agent Verification," arXiv preprint arXiv:2601.07504, Jan. 2026. [9]. Responsible AI Foundation, "MLE-Star: A Multi-Agent System for Machine Learning Engineering," Aug. 2025. [Online]. Available: https://www.responsibleaifoundation.com/post/mle-star-a-multi-agent-system-for-machine-learning-engineering [10]. IEEE Xplore, "ASPIRE: A Multi-Agent Framework for Execution-Free Code Analysis and Repair," in Proc. IEEE International Conference on Big Data, Dec. 2024. doi: 10.1109/BigData.2024.10825553 [11]. J. He, C. Treude, and D. Lo, "LLM-based multi-agent systems for software engineering: Literature review, vision and the road ahead," 2024. [12]. C. Lee et al., "A unified debugging approach via llm-based multi-agent synergy," 2024. [13]. X. Tang et al., "Codeagent: Autonomous communicative agents for code review," 2024. [14]. L. Wang et al., "Unity is strength: Collaborative llm-based agents for code reviewer recommendation," Oct. 2024, pp. 2235–2239. [15]. I. Ozkaya, "Application of large language models to software engineering tasks: Opportunities, risks, and implications," IEEE Software, vol. 40, no. 3, pp. 4–8, 2023. [16]. N. Johnson et al., "Static analysis tool evaluation on real-world codebases," 2024. [17]. Y. Wang and J. Zhu, "Meta Prompt Testing for metamorphic code verification," 2024. [18]. Dilgren et al., "SecRepoBench: Security benchmark for LLM-generated code," 2025. [19]. Vero et al., "BaxBench: Backend code vulnerability detection," 2025. [20]. Jimenez et al., "SWE-bench: Evaluating LLMs on real-world GitHub issues," 2024. |
