LoRKD: Low-Rank Knowledge Decomposition for Medical Foundation Models

TL;DR

LoRKD introduces a low-rank knowledge decomposition framework that enhances medical foundation models by improving task-specific performance and efficiency through expert modules and knowledge separation convolution.

Contribution

The paper proposes a novel Low-Rank Knowledge Decomposition framework that explicitly separates gradients for better specialization and efficiency in medical foundation models.

Findings

Achieves state-of-the-art performance on segmentation and classification tasks.

Demonstrates superior transferability on downstream tasks.

Reduces resource consumption while maintaining high accuracy.

Abstract

The widespread adoption of large-scale pre-training techniques has significantly advanced the development of medical foundation models, enabling them to serve as versatile tools across a broad range of medical tasks. However, despite their strong generalization capabilities, medical foundation models pre-trained on large-scale datasets tend to suffer from domain gaps between heterogeneous data, leading to suboptimal performance on specific tasks compared to specialist models, as evidenced by previous studies. In this paper, we explore a new perspective called "Knowledge Decomposition" to improve the performance on specific medical tasks, which deconstructs the foundation model into multiple lightweight expert models, each dedicated to a particular anatomical region, with the aim of enhancing specialization and simultaneously reducing resource consumption. To accomplish the above objective, we propose a novel framework named Low-Rank Knowledge Decomposition (LoRKD), which explicitly separates gradients from different tasks by incorporating low-rank expert modules and efficient knowledge separation convolution. The low-rank expert modules resolve gradient conflicts between heterogeneous data from different anatomical regions, providing strong specialization at lower costs. The efficient knowledge separation convolution significantly improves algorithm efficiency by achieving knowledge separation within a single forward propagation. Extensive experimental results on segmentation and classification tasks demonstrate that our decomposed models not only achieve state-of-the-art performance but also exhibit superior transferability on downstream tasks, even surpassing the original foundation models in task-specific evaluations. The code is available at here.