Research map / Label-efficient and robust learning / Semi-supervised learning
Active Learning on a Budget: Opposite Strategies Suit High and Low Budgets
Guy Hacohen, Avihu Dekel, Daphna Weinshall
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 |
|---|---|---|---|
| select diverse typical examples with TypiClust instead of random selection | FlexMatch · CIFAR-10, 10 labeled examples | accuracy: 39.4% | Improved |
| use TypiClust rather than random, balanced or unbalanced selection | FlexMatch · CIFAR-100, 300 labeled examples | accuracy | Improved |
| use TypiClust rather than random, balanced or unbalanced selection | FlexMatch · TinyImageNet, 1000 labeled examples | accuracy | Improved |
| use TypiClust instead of baseline active-learning strategies | fully supervised image classification · CIFAR-10, CIFAR-100, and ImageNet-100, low budget | accuracy | Improved |
| use TypiClust instead of baseline active-learning strategies | linear classifier on self-supervised embeddings · CIFAR-10, CIFAR-100, and ImageNet-100, low budget | accuracy | Improved |
| use random initial pool selection (ablation) | TypiClust · fully supervised low-budget experiments | accuracy | Hurt |
| select a random example from each cluster instead of the most typical example (ablation) | TypiClust · CIFAR-10, low budget | accuracy | Hurt |
| select the most atypical example in each cluster instead of the most typical (ablation) | TypiClust · CIFAR-10, low budget | accuracy | Hurt |
| remove clustering and select typical samples greedily (ablation) | TypiClust · CIFAR-10, low budget | accuracy | Hurt |
| use oracle-estimated uncertainty to query examples (ablation) | active learning on CIFAR-10 · CIFAR-10, low budget | accuracy | Hurt |
| prefer typical examples rather than uncertain examples | active learning · low-budget regime | generalization error | Improved |
| prefer atypical or confusing examples rather than typical examples | active learning · high-budget regime | generalization error | Improved |
Closest papers by meaning
- Active Learning Through a Covering Lens (NeurIPS 2022 · similarity 0.87)
- A Cross-Domain Benchmark for Active Learning (NeurIPS 2024 · similarity 0.85)
- Towards Cost-Effective Learning: A Synergy of Semi-Supervised and Active Learning (CVPR 2025 · similarity 0.85)
- Enhancing Semi-Supervised Learning via Representative and Diverse Sample Selection (NeurIPS 2024 · similarity 0.84)
- Active Learning with LLMs for Partially Observed and Cost-Aware Scenarios (NeurIPS 2024 · similarity 0.84)
- Contextual Active Model Selection (NeurIPS 2024 · similarity 0.83)
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.
