AI Paper Atlas

Research map / Continual learning / Continual fine-tuning of LLMs

Forging a Dynamic Memory: Retrieval-Guided Continual Learning for Generalist Medical Foundation Models

Zizhi Chen, Yizhen Gao, Minghao Han, Yizhou Liu, Zhaoyu Chen, Dingkang Yang, Lihua Zhang

CVPR 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Mixture-of-experts for multimodal continual tuning · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use last-task model as distillation teacher instead of initial CLIP (ablation)PRIMED · HieraMedTransfer Order ILast: 78.2 -> 82.1Improved
Use contrastive (image+text) distillation instead of image-only (ablation)PRIMED · HieraMedTransfer Order ILast: 81.3 -> 82.1Improved
Use contrastive distillation instead of text-only (ablation)PRIMED · HieraMedTransfer Order ILast: 80.5 -> 82.1Improved
Use Dynamic Fisher Weight Guard instead of l2 (ablation)PRIMED · HieraMedTransfer Order ITransfer: 57.2 -> 58.3Improved
Use Dynamic Fisher Weight Guard instead of EWC (ablation)PRIMED · HieraMedTransfer Order IAvg: 71.3 -> 73.1Improved
Use hierarchical retrieval instead of BM25 (ablation)PRIMED · HieraMedTransfer Order ITransfer: 55.1 -> 58.3Improved
Use hierarchical retrieval instead of embedding-only (ablation)PRIMED · HieraMedTransfer Order ITransfer: 56.1 -> 58.3Improved
Add CMC loss (beta 0 -> 0.25) (ablation)PRIMED · HieraMedTransfer Order ITransfer / Last: Transfer 57.2 -> 58.3; Last 82.5 -> 82.1Mixed
Use WISE-weighted teacher instead of last-task teacher (ablation)PRIMED · HieraMedTransfer Order ILast: 82.1 -> 79.1Hurt
Dynamic multi-stage retrieval reference data instead of uniform ZSCL reference dataPRIMED · HieraMedTransfer Order IILast: 77.0 -> 81.2Improved
Dynamic retrieval vs uniform reference dataPRIMED · MedXtreme Order IIACC: 64.5 -> 68.1Improved
Apply LwF baselineLwF · HieraMedTransfer Order IAvg: 68.5 -> 54.3Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy