AI Paper Atlas

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

Dynamic Multi-Layer Null Space Projection for Vision-Language Continual Learning

Borui Kang, Lei Wang, Zhiping Wu, Tao Feng, Yawen Li, Yang Gao, Wenbin Li

ICCV 2025 · Continual learning › Continual fine-tuning of LLMs · approach: Mixture-of-experts for multimodal continual tuning · 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
apply multi-layer null space projection to vision branch only with dynamic lambda (ablation)Adapter-based CLIP · TinyImageNet CIL (5 subsets)Last accuracy: 78.24 -> 79.14Improved
apply multi-layer projection to both modalities (ablation)Adapter-based CLIP · TinyImageNet CIL (5 subsets)Last accuracy: 78.24 -> 78.12No clear effect
apply DMNSP to language branch only (ablation)Adapter-based CLIP · TinyImageNet CIL (5 subsets)Last accuracy: 78.24 -> 77.22Hurt
apply single-layer projection to vision only plus lambda (ablation)Adapter-based CLIP · TinyImageNet CIL (5 subsets)Last accuracy: 78.24 -> 78.46No clear effect
add traditional single-layer null space projection (GPM) to adapter CLIPAdapter-based CLIP · TinyImageNet CIL (10 subsets)Last forgetting rate: 8.75 -> 8.69No clear effect
add DMNSP to adapter CLIPAdapter-based CLIP · TinyImageNet CIL (10 subsets)Last forgetting rate: 8.75 -> 4.84Improved
add DMNSP to MoE4AdaptersMoE4Adapters · TinyImageNet CIL (5 subsets)Average accuracy: 81.12 -> 81.19Improved
use simpler adapter with DMNSP instead of MoE4AdaptersCLIP continual learning · CIFAR100 training costtime per iteration: 1.58s/it -> 0.23s/itImproved
DMNSP vs TRGP on CIFAR100 50 subsets (average)Adapter-based CLIP · CIFAR100 CIL (50 subsets)Average accuracy: 83.70 -> 83.66No clear effect

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