UniPrompt-CL: Sustainable Continual Learning in Medical AI with Unified Prompt Pools

Researchers developed UniPrompt-CL, a new continual learning method specifically designed for medical AI that addresses the limitations of existing approaches when applied to medical data. The method uses a unified prompt pool design and regularization to achieve better performance while reducing computational costs, improving accuracy by 1-3 percentage points in domain-incremental learning settings.

• →UniPrompt-CL is a medical-oriented continual learning method that addresses domain bias and institutional constraints in medical AI.
• →The approach uses a minimally expanding unified prompt pool with new regularization terms for better stability-plasticity trade-off.
• →The method reduces inference costs while improving average accuracy by 1-3 percentage points across two domain-incremental learning settings.
• →Traditional continual learning methods designed for natural images often fail to transfer effectively to medical data.
• →The research validates that specialized approaches are needed for continual learning in medical AI applications.