Research map / Vision-language and multimodal / Cross-modal learning
Multi-Modal Representation Learning via Semi-Supervised Rate Reduction for Generalized Category Discovery
Wei He, Xianghan Meng, Zhiyuan Huang, Xianbiao Qi, Rong Xiao, Chun-Guang Li
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 dual-branch co-taught classifiers (ablation) | single shared classifier baseline · Stanford Cars | ACC (All): 75.2 -> 81.7 | Improved |
| add dual-branch co-taught classifiers (ablation) | single shared classifier baseline · Flowers102 | ACC (All): 78.3 -> 83.9 | Improved |
| add Retrieval-based Text Aggregation (RTA) (ablation) | dual-branch classifiers · Stanford Cars | ACC (All): 81.7 -> 86.0 | Improved |
| add Retrieval-based Text Aggregation (RTA) (ablation) | dual-branch classifiers · Flowers102 | ACC (All): 83.9 -> 87.4 | Improved |
| add L_SSR2 representation loss (ablation) | dual-branch classifiers + RTA · Stanford Cars | ACC (All): 86.0 -> 89.2 | Improved |
| add L_SSR2 representation loss (ablation) | dual-branch classifiers + RTA · Flowers102 | ACC (All): 87.4 -> 93.5 | Improved |
| add L_SSR2 representation loss (ablation) | dual-branch classifiers · Flowers102 | ACC (All): 83.9 -> 89.1 | Improved |
| add inter-modal L_CLIP to L_SSR2 (ablation) | SSR2-GCD · Stanford Cars | ACC (All): 89.2 -> 88.1 | Hurt |
| add inter-modal L_CLIP to L_SSR2 (ablation) | SSR2-GCD · Flowers102 | ACC (All): 93.5 -> 92.9 | Hurt |
| add inter-modal L_CLIP to L_SSR2 with low-quality prompts (CLIP-B/16) (ablation) | SSR2-GCD · Stanford Cars | ACC (All): 85.2 -> 84.0 | Hurt |
| add inter-modal L_CLIP to contrastive L_con (ablation) | SSR2-GCD framework with L_con · Stanford Cars | ACC (All): 87.9 -> 86.7 | Hurt |
| replace L_con with L_SSR2 (ablation) | SSR2-GCD framework · Flowers102 | ACC (All): 91.8 -> 93.5 | Improved |
Closest papers by meaning
- GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery (CVPR 2025 · similarity 0.88)
- Consistent Supervised-Unsupervised Alignment for Generalized Category Discovery (NeurIPS 2025 · similarity 0.87)
- SEAL: Semantic-Aware Hierarchical Learning for Generalized Category Discovery (NeurIPS 2025 · similarity 0.87)
- Transfer and Alignment Network for Generalized Category Discovery (AAAI 2024 · similarity 0.86)
- Less Attention is More: Prompt Transformer for Generalized Category Discovery (CVPR 2025 · similarity 0.86)
- HiLo: A Learning Framework for Generalized Category Discovery Robust to Domain Shifts (ICLR 2025 · 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.
