Research map / Label-efficient and robust learning / Semi-supervised learning
PromptCAL: Contrastive Affinity Learning via Auxiliary Prompts for Generalized Novel Category Discovery
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 PromptCAL instead of the strongest listed baseline | CIFAR-10 category discovery · CIFAR-10 | accuracy (All): 96.9 -> 97.9 | Helped |
| Use PromptCAL instead of the strongest listed baseline | CIFAR-100 category discovery · CIFAR-100 | accuracy (All): 74.2 -> 81.2 | Helped |
| Use PromptCAL instead of the strongest listed baseline | ImageNet-100 category discovery · ImageNet-100 | accuracy (All): 79.2 -> 83.1 | Helped |
| Use PromptCAL instead of the strongest listed baseline | CUB-200 category discovery · CUB-200 | accuracy (All): 51.3 -> 62.9 | Helped |
| Use PromptCAL instead of the strongest listed baseline | StanfordCars category discovery · StanfordCars | accuracy (All): 39.0 -> 50.2 | Helped |
| Use PromptCAL instead of the strongest listed baseline | Aircraft category discovery · Aircraft | accuracy (All): 45.0 -> 52.2 | Helped |
| Add affinity propagation to consensus KNN and SemiPriori (ablation) | PromptCAL · CUB-200 | accuracy (All): 54.6 -> 62.9 | Helped |
| Add SemiPriori pairwise label constraints (ablation) | PromptCAL · CUB-200 | accuracy (All): 57.3 -> 62.9 | Helped |
| Retain SemiCL alongside CAL (ablation) | PromptCAL · CUB-200 | accuracy (All): 61.7 -> 62.9 | Helped |
| Use SemiAG instead of naive KNN with SemiPriori (ablation) | PromptCAL · CUB-200 | accuracy (All): 60.1 -> 62.9 | Helped |
| Add prompt-related components (ablation) | PromptCAL · CUB-200 | accuracy (All): 60.3 -> 62.9 | Helped |
| Add prompt SemiCL and prompt CAL losses (ablation) | PromptCAL · CUB-200 | accuracy (All): 51.1 -> 62.9 | Helped |
Closest papers by meaning
- SPTNet: An Efficient Alternative Framework for Generalized Category Discovery with Spatial Prompt Tuning (ICLR 2024 · similarity 0.89)
- Less Attention is More: Prompt Transformer for Generalized Category Discovery (CVPR 2025 · similarity 0.89)
- Beyond Words: Augmenting Discriminative Richness via Diffusions in Unsupervised Prompt Learning (CVPR 2025 · similarity 0.87)
- Concept-Guided Prompt Learning for Generalization in Vision-Language Models (AAAI 2024 · similarity 0.86)
- Semantic-Guided Novel Category Discovery (AAAI 2024 · similarity 0.86)
- SEAL: Semantic-Aware Hierarchical Learning for Generalized Category Discovery (NeurIPS 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.