UniPrompt-CL：基于统一提示池的医疗AI可持续持续学习
UniPrompt-CL: Sustainable Continual Learning in Medical AI with Unified Prompt Pools
AI总结：
UniPrompt-CL通过改进提示池设计和引入新正则化项，在医疗领域实现更稳定的持续学习，降低计算成本并提升模型性能。
摘要：
现代AI模型通常在静态数据集上训练，限制了其适应快速变化的现实环境的能力。虽然持续学习（CL）解决了这一限制，但大多数CL方法针对自然图像设计，难以在医疗数据上有效迁移，因存在领域偏差、机构限制和细微的阶段边界。我们提出UniPrompt-CL，一种面向医疗的基于提示的持续学习方法，通过最小扩展的统一提示池和新正则化项改进提示池设计，实现更好的稳定性-可塑性平衡，计算成本更低。在两个领域增量学习设置中，UniPrompt-CL有效降低了推理成本，同时将AvgACC提升1-3个百分点。除了强大性能外，广泛实验清晰验证了所提改进的动机和有效性。
英文摘要：
Modern AI models are typically trained on static datasets, limiting their ability to continuously adapt to rapidly evolving real-world environments. While continual learning (CL) addresses this limitation, most CL methods are designed for natural images and often underperform or fail to transfer to medical data due to domain bias, institutional constraints, and subtle inter-stage boundaries. We propose UniPrompt-CL, a medical-oriented prompt-based continual learning method that improves prompt pool design via a minimally expanding unified prompt pool and a new regularization term, achieving a better stability-plasticity trade-off with lower computational cost. Across two domain-incremental learning settings, UniPrompt-CL effectively reduces inference cost while improving AvgACC by 1-3 percentage points. In addition to strong performance, extensive experiments clearly validate the motivation and effectiveness of the proposed improvements.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
