AI Paper Atlas

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

SAMCL: Empowering SAM to Continually Learn from Dynamic Domains with Extreme Storage Efficiency

Zeqing Wang, Kangye Ji, Di Wang, Haibin Zhang, Fei Cheng

AAAI 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Low-rank and model-merging adaptation · 4 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
add AugModule (SLoRA + prompt augmentation) to LoRA baseline (ablation)SAMCL (SAM) · CL Kvasir to CODAA mIoU: 0.703 -> 0.748Improved
add Module Selector on top of AugModule (ablation)SAMCL (SAM) · CL Kvasir to CODAA mIoU: 0.748 -> 0.836Improved
add Module Selector on top of AugModule (ablation)SAMCL (SAM) · CL Kvasir to CODFM mIoU: 0.086 -> 0.0019Improved
add Module Selector on top of AugModule (ablation)SAMCL (SAM) · CL Kvasir to CODFT mIoU: 0.570 -> 0.668Improved
add prompt augmentation (AugModule instead of SLoRA) (ablation)SLoRA (SAM) · CODmIoU: 0.735 -> 0.782Improved
share matrix A across layers (SLoRA instead of LoRA) (ablation)LoRA (SAM) · CODmIoU: 0.733 -> 0.735No clear effect
share matrix A across layers (SLoRA instead of LoRA) (ablation)LoRA (SAM) · Parameter storageMB: 0.308 -> 0.190Improved
store 300 embeddings per task instead of 10 (ablation)SAMCL (SAM) · CL Kvasir to CODFM mIoU: -0.0546 -> 0.0009Improved
store 500 embeddings per task instead of 10 (ablation)SAMCL (SAM) · CL Kvasir to CODAA mIoU: 0.808 -> 0.838Improved
train Module Selector on COCO subset after CL (ablation)SAMCL (SAM) · Revert to original SAMmIoU: 0.5281 -> 0.6162Improved
use SAMCL instead of MoDAMoDA (SAM) · CL Kvasir to CODAA mIoU: 0.756 -> 0.836Improved
use SAMCL instead of ER (replay)ER (SAM) · CL Kvasir to CODFT mIoU: 0.630 -> 0.668Improved

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