Low-Rank Knowledge Decomposition for Medical Foundation Models

arXiv:2404.17184v1 Announce Type: new
Abstract: The popularity of large-scale pre-training has promoted the development of medical foundation models. However, some studies have shown that although foundation models exhibit strong general feature extraction capabilities, their performance on specific tasks is still inferior to task-specific methods. In this paper, we explore a new perspective called ``Knowledge Decomposition'' to improve the performance on specific medical tasks, which deconstruct the foundation model into multiple lightweight expert models, each dedicated to a particular task, with …

abstract arxiv capabilities cs.cv development explore extraction feature feature extraction foundation general however knowledge low medical paper performance perspective pre-training promoted scale specific tasks studies tasks training type