Research map / Vision-language and multimodal / CLIP and zero-shot
Enhancing Few-Shot Vision-Language Classification with Large Multimodal Model Features
Chancharik Mitra, Brandon Huang, Tianning Chai, Zhiqiu Lin, Assaf Arbelle, Rogerio Feris, Leonid Karlinsky, Trevor Darrell, Deva Ramanan, Roei Herzig
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 |
|---|---|---|---|
| use class centroid classification instead of KNN for head selection (ablation) | SAVs (LLaVA-OneVision-7B) · MHaluBench | accuracy: 53.0 -> 80.8 | Improved |
| use class centroid instead of linear probing (ablation) | SAVs (LLaVA-OneVision-7B) · MHaluBench / NaturalBench / EuroSAT | accuracy: MHB 82.5 vs 80.8; NB 32.9 vs 35.1; ES 83.1 vs 86.7 | Mixed |
| select sparse heads instead of sparse layers (ablation) | SAVs (LLaVA-OneVision-7B) · NaturalBench (group score) | accuracy: 28.4 -> 35.1 | Improved |
| select sparse heads instead of sparse layers (ablation) | SAVs (LLaVA-OneVision-7B) · EuroSAT | accuracy: 81.8 -> 86.7 | Improved |
| add 2 or 5 noisy examples per class (ablation) | SAVs (LLaVA-OneVision-7B) · NaturalBench (group score) | accuracy: 35.1 -> 36.1 | No clear effect |
| add 10 noisy examples per class out of 20 (ablation) | SAVs (LLaVA-OneVision-7B) · MHaluBench | accuracy: 82.0 -> 50.3 | Hurt |
| extract from last token instead of first token (ablation) | SAVs (LLaVA-OneVision-7B) · EuroSAT | accuracy: 24.9 -> 86.7 | Improved |
| extract from last token instead of middle token (ablation) | SAVs (LLaVA-OneVision-7B) · NaturalBench (group score) | accuracy: 2.4 -> 35.1 | Improved |
| add online learning with randomized weighted voting (ablation) | SAVs (LLaVA-OneVision-7B) · MHaluBench | accuracy: 82.0 -> 73.2 | Hurt |
| use 20 one-shot examples instead of formatted ICL input (ablation) | SAVs (LLaVA-OneVision-7B) · MHaluBench | accuracy: 28.3 -> 82.0 | Improved |
| add SAVs to the zero-shot LMM | SAVs (LLaVA-OneVision-7B) · VLGuard | accuracy: 31.4 -> 94.3 | Improved |
| add 4-shot ICL to the zero-shot LMM | SAVs (LLaVA-OneVision-7B) · MHaluBench | accuracy: 34.7 -> 25.0 | Hurt |
Closest papers by meaning
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 2025 · similarity 0.90)
- Large Language Models are Good Prompt Learners for Low-Shot Image Classification (CVPR 2024 · similarity 0.90)
- VladVA: Discriminative Fine-tuning of LVLMs (CVPR 2025 · similarity 0.90)
- MMRL: Multi-Modal Representation Learning for Vision-Language Models (CVPR 2025 · similarity 0.89)
- Unified Generative and Discriminative Training for Multi-modal Large Language Models (NeurIPS 2024 · similarity 0.89)
- Rethinking Few-Shot Adaptation of Vision-Language Models in Two Stages (CVPR 2025 · similarity 0.89)
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.
