AI Paper Atlas

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

DSCA: Dynamic Subspace Concept Alignment for Lifelong VLM Editing

Gyanendra Das, Sai Jena

CVPR 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Mixture-of-experts for multimodal continual tuning · 0 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 orthogonality constraint in subspaces (ablation)DSCA · AblationEdit Success: 95.8 -> 98.0Improved
add orthogonality constraint in subspaces (ablation)DSCA · AblationLocality drop: 2.8 -> 0.5Improved
add orthogonality constraint in subspaces (ablation)DSCA · AblationGeneralization: 93.4 -> 97.3Improved
add gate sparsity loss (ablation)DSCA · AblationEdit Success: 96.1 -> 98.0Improved
add gate sparsity loss (ablation)DSCA · AblationLocality drop: 2.1 -> 0.5Improved
use two-stage hierarchical routing instead of single stage (ablation)DSCA · AblationLocality drop: 1.9 -> 0.5Improved
add basis-residual intervention (ablation)DSCA · AblationLocality drop: 1.5 -> 0.5Improved
use full number of subspaces K instead of K/2 (ablation)DSCA · AblationEdit Success: 97.0 -> 98.0Improved
use full rank r instead of r/2 (ablation)DSCA · AblationLocality drop: 1.2 -> 0.5Improved
residualized incremental PCA vs no orthogonalization (ablation)DSCA · 1000 sequential editssubspace overlap: 10^-1 -> 3x10^-3Improved
apply DSCA instead of LiveEdit in lifelong editingDSCA · E-VQA lifelong (1000 edits)Rel.: 92.93 -> 96.85Improved
apply DSCA instead of PAM on CoINDSCA · CoINBWT: -19.45 -> -9.37Improved

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