ReclAIm: A multi-agent framework for degradation-aware performance tuning of medical imaging AI

2025-10-19

AI Summary

To address performance degradation of AI models for medical imaging in clinical deployment due to data drift, this paper proposes a large language model (LLM)-driven multi-agent collaborative framework enabling fully automated, programming-free continuous monitoring, evaluation, and adaptive fine-tuning. The framework comprises three specialized agents—monitoring, evaluation, and fine-tuning—orchestrated by an LLM acting as a central controller; they interact via natural language to close the optimization loop and support cross-modal imaging modalities (e.g., MRI, CT, X-ray). Its key innovation lies in unifying transfer learning and parameter-efficient fine-tuning within an LLM-mediated end-to-end autonomous maintenance pipeline. Experiments demonstrate that the method restores degraded model performance to within ±1.5% of its original level, achieving a maximum performance recovery of 41.1%, thereby significantly enhancing long-term clinical reliability.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Ensuring the long-term reliability of AI models in clinical practice requires continuous performance monitoring and corrective actions when degradation occurs. Addressing this need, this manuscript presents ReclAIm, a multi-agent framework capable of autonomously monitoring, evaluating, and fine-tuning medical image classification models. The system, built on a large language model core, operates entirely through natural language interaction, eliminating the need for programming expertise. ReclAIm successfully trains, evaluates, and maintains consistent performance of models across MRI, CT, and X-ray datasets. Once ReclAIm detects significant performance degradation, it autonomously executes state-of-the-art fine-tuning procedures that substantially reduce the performance gap. In cases with performance drops of up to -41.1% (MRI InceptionV3), ReclAIm managed to readjust performance metrics within 1.5% of the initial model results. ReclAIm enables automated, continuous maintenance of medical imaging AI models in a user-friendly and adaptable manner that facilitates broader adoption in both research and clinical environments.

Problem

• Autonomously monitors and fine-tunes medical imaging AI models for degradation
• Maintains consistent performance across MRI, CT, and X-ray datasets
• Enables automated continuous maintenance through natural language interaction

Innovation

• Multi-agent framework monitors medical imaging AI degradation
• Natural language interaction eliminates programming expertise requirement
• Autonomous fine-tuning reduces performance gap after detection