Research map / Label-efficient and robust learning / Recognition and long-tail
CloSeR: Unified Relational Distillation from Closed-Set Teachers for Category Discovery
Yuanpei Liu, Zhenqi He, Jialu Tang, Kai Han
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 CloSeR (closed-set teacher + URD) | SimGCD · Stanford-Cars (DINO) | ACC All (%): 53.8 -> 63.0 | Improved |
| add CloSeR | SelEx · CUB (DINO) | ACC All (%): 73.6 -> 82.8 | Improved |
| add CloSeR | SimGCD · fine-grained average (DINOv2) | ACC New (%): 65.3 -> 74.5 | Improved |
| add CloSeR | SimGCD · generic average (DINO) | ACC All (%): 86.7 -> 88.8 | Improved |
| add CloSeR | SimGCD · ImageNet-100 (DINOv2) | ACC All (%): 89.9 -> 92.5 | Improved |
| add global relation distillation (ablation) | SelEx · Stanford-Cars (DINO) | ACC All (%): 58.5 -> 61.9 | Improved |
| add local relation distillation (ablation) | SelEx · Stanford-Cars (DINO) | ACC All (%): 58.5 -> 64.3 | Improved |
| combine global and local relations (ablation) | SelEx · FGVC-Aircraft (DINO) | ACC All (%): 59.0/63.0 -> 64.9 | Improved |
| add feature decoupling between global and local pathways (ablation) | SelEx + URD · FGVC-Aircraft (DINO) | ACC All (%): 64.9 -> 66.6 | Improved |
| add feature decoupling (ablation) | SelEx + URD · Stanford-Cars (DINO) | ACC All (%): 65.8 -> 66.5 | Improved |
| use block-wise adapter instead of fine-tuning only last block for the teacher (ablation) | CloSeR teacher · Stanford-Cars (DINO) | ACC All (%): 46.7 -> 63.0 | Improved |
| use block-wise adapter instead of full fine-tuning for the teacher (ablation) | CloSeR teacher · CUB (DINO) | ACC All (%): 68.1 -> 68.2 | No clear effect |
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.
- SelEx — SelEx: Self-Expertise in Fine-Grained Generalized Category Discovery (ECCV 2024) · better on 3 of 3 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
- A Unified Knowledge Transfer Network for Generalized Category Discovery (AAAI 2024 · similarity 0.90)
- Transfer and Alignment Network for Generalized Category Discovery (AAAI 2024 · similarity 0.90)
- Towards Understanding Parametric Generalized Category Discovery on Graphs (ICML 2025 · similarity 0.89)
- Hyperbolic Category Discovery (CVPR 2025 · similarity 0.89)
- HiLo: A Learning Framework for Generalized Category Discovery Robust to Domain Shifts (ICLR 2025 · similarity 0.89)
- Prior-Constrained Association Learning for Fine-Grained Generalized Category Discovery (AAAI 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.
