Research map / Label-efficient and robust learning / Recognition and long-tail
LLM-assisted Entropy-based Adaptive Distillation for Unsupervised Fine-grained Visual Representation Learning
Jianfeng Dong, Danfeng Luo, Daizong Liu, Jie Sun, Xiaoye Qu, Xun Yang, Dongsheng Liu, Xun 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 |
|---|---|---|---|
| add CLIP knowledge distillation to MoCo (ablation) | MoCo · Cub classification | Top-1: 62.91 -> 75.46 | Improved |
| add contrastive learning to distillation (ablation) | LEAD · Cub classification | Top-1: 75.46 -> 76.03 | Improved |
| add entropy-based adaptive weighting (ablation) | LEAD (CL+KD) · Cub classification | Top-1: 76.03 -> 78.44 | Improved |
| add entropy-based adaptive weighting (ablation) | LEAD (CL+KD) · Cub retrieval | mAP: 38.87 -> 44.43 | Improved |
| use LLM descriptions instead of fixed template (ablation) | LEAD · Cub classification | Top-1: 76.63 -> 78.44 | Improved |
| use LLM descriptions instead of fixed template (ablation) | LEAD · Car classification | Top-1: 85.94 -> 88.14 | Improved |
| use LLM descriptions instead of fixed template (ablation) | LEAD · Aircraft classification | Top-1: 64.81 -> 67.45 | Improved |
| use LLM-assisted distillation instead of feature-based distillation (ablation) | LEAD · Cub retrieval | mAP: 28.64 -> 44.43 | Improved |
| use LLM-assisted distillation instead of relation-based distillation (ablation) | LEAD · Car classification | Top-1: 75.68 -> 88.14 | Improved |
| use information entropy instead of maximum probability for confidence (ablation) | LEAD · Aircraft classification | Top-1: 63.67 -> 67.45 | Improved |
| use information entropy instead of maximum probability for confidence (ablation) | LEAD · Cub classification | Top-1: 78.07 -> 78.44 | Improved |
| replace dataset categories with LLM-generated categories (ablation) | LEAD · Cub classification | Top-1: 78.44 -> 78.20 | No clear effect |
Closest papers by meaning
- Democratizing Fine-grained Visual Recognition with Large Language Models (ICLR 2024 · similarity 0.89)
- Fine-R1: Make Multi-modal LLMs Excel in Fine-Grained Visual Recognition by Chain-of-Thought Reasoning (ICLR 2026 · similarity 0.89)
- Analyzing and Boosting the Power of Fine-Grained Visual Recognition for Multi-modal Large Language Models (ICLR 2025 · similarity 0.87)
- GranViT: A Fine-Grained Vision Model With Autoregressive Perception For MLLMs (ICLR 2026 · similarity 0.86)
- DiVE-k: DIFFERENTIAL VISUAL REASONING FOR FINE-GRAINED IMAGE RECOGNITION (ICLR 2026 · similarity 0.86)
- Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation (ICLR 2026 · similarity 0.86)
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.
