Research map / Label-efficient and robust learning / Semi-supervised learning
PSS: Progressive Sample Selection for Open-World Visual Representation Learning
Tianyue Cao, Yongxin Wang, Yifan Xing, Tianjun Xiao, Tong He, Zheng Zhang, Hao Zhou, Joseph Tighe
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 sample selection to iterative training (ablation) | PSS · iNaturalist retrieval | Recall@1: 0.5548 -> 0.5714 | Improved |
| add iterative refinement to sample selection (ablation) | PSS · iNaturalist retrieval | Recall@1: 0.5522 -> 0.5714 | Improved |
| add sample selection (fewer training images) (ablation) | PSS · iNaturalist retrieval | training data: 325,846 -> 87,349 | Improved |
| add sample selection with one-step training (ablation) | Hi-LANDER pseudo-labelling · iNaturalist retrieval | Recall@1: 0.5421 -> 0.5522 | Improved |
| train on selected samples far from labeled classes instead of close ones (ablation) | PSS · iNaturalist retrieval (iteration 1) | Recall@1: 0.5490 -> 0.5524 | Improved |
| add progressive selection instead of all-data iterations (DeepClustering) | PSS · iNaturalist retrieval | Recall@1: 0.5548 -> 0.5714 | Improved |
| add progressive selection instead of DeepClustering | PSS · IJB-C face verification | FNMR@FMR 1e-4: 0.2234 -> 0.2165 | Improved |
| add open-set UNO pseudo labels to supervised baseline | UNO · iNaturalist retrieval | Recall@1: 0.5376 -> 0.5372 | Hurt |
| add pseudo labeling (all unlabeled data at once) | Supervised baseline · iNaturalist retrieval | Recall@1: 0.5376 -> 0.5447 | Improved |
Closest papers by meaning
- Benchmarking Representation Learning for Natural World Image Collections (CVPR 2021 · similarity 0.88)
- Detecting Open World Objects via Partial Attribute Assignment (CVPR 2025 · similarity 0.85)
- Cross-domain Open-world Discovery (ICML 2024 · similarity 0.84)
- Open-World Objectness Modeling Unifies Novel Object Detection (CVPR 2025 · similarity 0.84)
- Targeted Representation Alignment for Open-World Semi-Supervised Learning (CVPR 2024 · similarity 0.84)
- VL-SAM-V2: Open-World Object Detection with General and Specific Query Fusion (NeurIPS 2025 · similarity 0.84)
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.
