AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Leveraging Cross-Modal Knowledge Transfer for Knowledge-Aware Concept Customization

chenyang zhu, Hongxiang Li, Xiu Li, Long Chen

ECCV 2026 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment · 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
update knowledge in a frozen LLM encoder after concept learning instead of finetuning the encoder (Enc-FT)MoKus · KnowCusBench generationCLIP-I-Seg: 0.562 -> 0.718Improved
update knowledge in a frozen LLM encoder instead of finetuning the encoder (Enc-FT)MoKus · KnowCusBench reconstructionCLIP-I-Seg: 0.553 -> 0.764Improved
closed-form knowledge update instead of per-knowledge DreamBooth retraining (Naive-DB)MoKus · KnowCusBench generationCLIP-T: 0.291 -> 0.305Improved
closed-form knowledge update instead of Naive-DBMoKus · KnowCusBench trainingtime (min): ~27min -> ~6minImproved
MoKus versus Naive-DB on CLIP-IMoKus · KnowCusBench reconstructionCLIP-I: 0.874 -> 0.867Hurt
MoKus versus Naive-DB on generation CLIP-IMoKus · KnowCusBench generationCLIP-I: 0.789 -> 0.761Hurt
scaling factor eta 1e-4 instead of 1e-6 (ablation)MoKus · KnowCusBench generationPick Score: 21.30 -> 18.40Hurt
scaling factor eta 1e-5 instead of 1e-6 (ablation)MoKus · KnowCusBench generationCLIP-I-Seg: 0.718 -> 0.717No clear effect
increase number of knowledge items from 1 to 5 (ablation)MoKus · KnowCusBench generationCLIP-T: 0.304 -> 0.305No clear effect
increase number of knowledge items from 1 to 5 (ablation)MoKus · KnowCusBench trainingtime (s): 331.3s -> 360.0sHurt
apply knowledge update with world knowledge to Qwen-ImageMoKus · WISE subsetWiScore: 0.81 -> 1.33Improved
apply knowledge update with world knowledge to Qwen-ImageMoKus · WISE subsetRealism: 0.70 -> 1.42Improved

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