{"ID":2840445,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.13099","arxiv_id":"2511.13099","title":"MergeSlide: Continual Model Merging and Task-to-Class Prompt-Aligned Inference for Lifelong Learning on Whole Slide Images","abstract":"Lifelong learning on Whole Slide Images (WSIs) aims to train or fine-tune a unified model sequentially on cancer-related tasks, reducing the resources and effort required for data transfer and processing, especially given the gigabyte-scale size of WSIs. In this paper, we introduce MergeSlide, a simple yet effective framework that treats lifelong learning as a model merging problem by leveraging a vision-language pathology foundation model. When a new task arrives, it is: 1) defined with class-aware prompts, 2) fine-tuned for a few epochs using an MLP-free backbone, and 3) merged into a unified model using an orthogonal continual merging strategy that preserves performance and mitigates catastrophic forgetting. For inference under the class-incremental learning (CLASS-IL) setting, where task identity is unknown, we introduce Task-to-Class Prompt-aligned (TCP) inference. Specifically, TCP first identifies the most relevant task using task-level prompts and then applies the corresponding class-aware prompts to generate predictions. To evaluate MergeSlide, we conduct experiments on a stream of six TCGA datasets. The results show that MergeSlide outperforms both rehearsal-based continual learning and vision-language zero-shot baselines. Code and data are available at https://github.com/caodoanh2001/MergeSlide.","short_abstract":"Lifelong learning on Whole Slide Images (WSIs) aims to train or fine-tune a unified model sequentially on cancer-related tasks, reducing the resources and effort required for data transfer and processing, especially given the gigabyte-scale size of WSIs. In this paper, we introduce MergeSlide, a simple yet effective fr...","url_abs":"https://arxiv.org/abs/2511.13099","url_pdf":"https://arxiv.org/pdf/2511.13099v1","authors":"[\"Doanh C. Bui\",\"Ba Hung Ngo\",\"Hoai Luan Pham\",\"Khang Nguyen\",\"Maï K. Nguyen\",\"Yasuhiko Nakashima\"]","published":"2025-11-17T07:51:18Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false,"code_links":[{"ID":606969,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2840445,"paper_url":"https://arxiv.org/abs/2511.13099","paper_title":"MergeSlide: Continual Model Merging and Task-to-Class Prompt-Aligned Inference for Lifelong Learning on Whole Slide Images","repo_url":"https://github.com/caodoanh2001/MergeSlide","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}