Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.10954 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915858147704832 |
|---|---|
| author | Oh, Gyutae Shin, Jitae |
| author_facet | Oh, Gyutae Shin, Jitae |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_10954 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | UniPrompt-CL: Sustainable Continual Learning in Medical AI with Unified Prompt Pools Oh, Gyutae Shin, Jitae Machine Learning Artificial Intelligence 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. |
| title | UniPrompt-CL: Sustainable Continual Learning in Medical AI with Unified Prompt Pools |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2508.10954 |