Research map / Vision-language and multimodal / Multimodal LLMs
Visual-Oriented Fine-Grained Knowledge Editing for MultiModal Large Language Models
Zhen Zeng, Leijiang Gu, Xun Yang, Zhangling Duan, Zenglin Shi, Meng Wang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| replace text-only scope classifier with multimodal scope classifier (ablation) | SERAC (BLIP-2 OPT) · FGVEdit | specificity (%): 31.92 -> 61.60 | Improved |
| replace text-only scope classifier with multimodal scope classifier (ablation) | SERAC (MiniGPT-4) · FGVEdit | specificity (%): 37.85 -> 57.20 | Improved |
| add multimodal scope classifier to MEND (MSCKE-MEND) (ablation) | MEND (BLIP-2 OPT) · FGVEdit | specificity (%): 65.85 -> 68.38 | Improved |
| add multimodal scope classifier to MEND (MSCKE-MEND) (ablation) | MEND (MiniGPT-4) · FGVEdit | specificity (%): 67.39 -> 71.98 | Improved |
| replace text-only scope classifier with multimodal scope classifier (ablation) | SERAC (BLIP-2 OPT) · FGVEdit | reliability (%): 93.08 -> 99.13 | Improved |
| replace text-only scope classifier with multimodal scope classifier (ablation) | SERAC (MiniGPT-4) · FGVEdit | reliability (%): 99.50 -> 99.50 | No clear effect |
| add multimodal scope classifier to MEND (ablation) | MEND (MiniGPT-4) · FGVEdit | reliability (%): 94.85 -> 97.05 | Improved |
| replace text-only scope classifier with multimodal scope classifier (ablation) | SERAC (BLIP-2 OPT) · FGVEdit | generality (%): 96.83 -> 98.56 | Improved |
| replace text-only scope classifier with multimodal scope classifier (ablation) | SERAC · MMEdit (VQA) | M-Locality (%): 2.91 -> 98.08 | Improved |
| replace text-only scope classifier with multimodal scope classifier (ablation) | SERAC · MMEdit (VQA) | T-Generality (%): 99.90 -> 95.60 | Hurt |
| add multimodal scope classifier to MEND (ablation) | MEND · MMEdit (VQA) | T-Generality (%): 97.51 -> 92.10 | Hurt |
| transfer classifier trained on BLIP-2 OPT to MiniGPT-4 instead of retraining (ablation) | MSCKE · FGVEdit | specificity (%): 57.09 -> 57.20 | No clear effect |
Closest papers by meaning
- MMKE-Bench: A Multimodal Editing Benchmark for Diverse Visual Knowledge (ICLR 2025 · similarity 0.94)
- ComprehendEdit: A Comprehensive Dataset and Evaluation Framework for Multimodal Knowledge Editing (AAAI 2025 · similarity 0.92)
- VLKEB: A Large Vision-Language Model Knowledge Editing Benchmark (NeurIPS 2024 · similarity 0.89)
- Towards Unified Multimodal Editing with Enhanced Knowledge Collaboration (NeurIPS 2024 · similarity 0.88)
- Matryoshka Multimodal Models (ICLR 2025 · similarity 0.88)
- LION: Empowering Multimodal Large Language Model with Dual-Level Visual Knowledge (CVPR 2024 · similarity 0.87)
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.
