Research map / Label-efficient and robust learning / Recognition and long-tail
CoGe-GCD: Reframing Generalized Category Discovery with Compositional Generalization
Luyao Tang, Jiewei Zheng, Kunze Huang, Chaoqi Chen, Yue Huang, Cheng Chen
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 column-normalized primitive competition (ablation) | CoGe-GCD (SimGCD) · CUB-200 | All accuracy: 62.6 -> 63.9 | Improved |
| add column-normalized primitive competition (ablation) | CoGe-GCD (SimGCD) · FGVC-Aircraft | All accuracy: 56.1 -> 57.9 | Improved |
| add column-normalized primitive competition (ablation) | CoGe-GCD (SimGCD) · CIFAR-100 | All accuracy: 78.5 -> 80.1 | Improved |
| add column-normalized primitive competition (ablation) | CoGe-GCD (SimGCD) · ImageNet-100 | All accuracy: 82.0 -> 83.0 | Improved |
| add geometric structure preserving calibration (ablation) | CoGe-GCD (SimGCD) · CUB-200 | All accuracy: 62.4 -> 63.9 | Improved |
| add geometric structure preserving calibration (ablation) | CoGe-GCD (SimGCD) · FGVC-Aircraft | All accuracy: 54.1 -> 57.9 | Improved |
| add geometric structure preserving calibration (ablation) | CoGe-GCD (SimGCD) · CIFAR-100 | All accuracy: 76.0 -> 80.1 | Improved |
| add geometric structure preserving calibration (ablation) | CoGe-GCD (SimGCD) · ImageNet-100 | All accuracy: 81.6 -> 83.0 | Improved |
| add CoGe-GCD plug-in to four GCD baselines | SimGCD, LegoGCD, CMS, SelEx · fine-grained average (CUB, Aircraft, Cars) | All accuracy: +4.03 | Improved |
| add CoGe-GCD plug-in to SimGCD | SimGCD · fine-grained CUB-200 Known accuracy | Known accuracy: 66.4 -> 65.5 | Hurt |
| add CoGe-GCD plug-in to baselines | SimGCD, LegoGCD, SimGCD on coarse · coarse-grained (CIFAR-10/100, ImageNet-100) | All accuracy: +1.07 | Improved |
| add CoGe-GCD to SimGCD on ImageNet-100 | SimGCD · ImageNet-100 | Known accuracy: 93.1 -> 92.4 | Hurt |
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
- Identifying Latent Concepts and Structures for Generalized Category Discovery (ICML 2026 · similarity 0.92)
- Hyperbolic Category Discovery (CVPR 2025 · similarity 0.89)
- Towards Understanding Parametric Generalized Category Discovery on Graphs (ICML 2025 · similarity 0.88)
- When Domain Generalization meets Generalized Category Discovery: An Adaptive Task-Arithmetic Driven Approach (CVPR 2025 · similarity 0.88)
- HIDISC: A Hyperbolic Framework for Domain Generalization with Generalized Category Discovery (NeurIPS 2025 · similarity 0.88)
- Federated Generalized Category Discovery (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.
