Research map / Label-efficient and robust learning / Recognition and long-tail
Influence Selection for Active Learning
Zhuoming Liu, Hao Ding, Huaping Zhong, Weijia Li, Jifeng Dai, Conghui He
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 |
|---|---|---|---|
| use top-1 pseudo-label gradient (K=1) instead of averaging top-10 classes (ablation) | ISAL · CIFAR10 (9000 labeled) | accuracy: 83.03 -> 89.26 | Improved |
| use K=1 instead of K=2 (ablation) | ISAL · CIFAR10 (3000 labeled) | accuracy: 65.65 -> 67.72 | Improved |
| use K=1 instead of K=5 (ablation) | ISAL · CIFAR10 (5000 labeled) | accuracy: 78.52 -> 81.24 | Improved |
| add inverse Hessian to the gradient-similarity influence (ablation) | ISAL · CIFAR10 (9000 labeled) | accuracy: 88.60 -> 89.26 | Improved |
| add inverse Hessian to the gradient-similarity influence (ablation) | ISAL · CIFAR10 (5000 labeled) | accuracy: 80.54 -> 81.24 | Improved |
| use validation set as reference set instead of initial labeled set L1 (ablation) | ISAL · CIFAR10 (5000 labeled) | accuracy: 80.57 -> 81.24 | Improved |
| use validation set as reference set instead of current labeled set (ablation) | ISAL · CIFAR10 (7000 labeled) | accuracy: 84.88 -> 85.96 | Improved |
| use labeled set as reference set versus random sampling (ablation) | ISAL_v3 · CIFAR10 (9000 labeled) | accuracy: 86.50 -> 88.71 | Improved |
| select samples by ISAL instead of coreset | ISAL · CIFAR10 | annotation saved: saving 12% of annotation | Improved |
| select samples by ISAL instead of random sampling | ISAL · CIFAR10 | annotation saved: saving 23% of annotation | Improved |
| select samples by ISAL instead of coreset | ISAL · CIFAR100 | annotation saved: saving 2.9% and 9.3% of annotation respectively | Improved |
| select samples by ISAL instead of coreset | ISAL · VOC2012 | annotation saved: saving 13% of annotations | Improved |
Closest papers by meaning
- Provably Neural Active Learning Succeeds via Prioritizing Perplexing Samples (ICML 2024 · similarity 0.86)
- Towards Cost-Effective Learning: A Synergy of Semi-Supervised and Active Learning (CVPR 2025 · similarity 0.85)
- Neural Networks for Learnable and Scalable Influence Estimation of Instruction Fine-Tuning Data (NeurIPS 2025 · similarity 0.85)
- Contextual Active Model Selection (NeurIPS 2024 · similarity 0.85)
- Enhancing Semi-Supervised Learning via Representative and Diverse Sample Selection (NeurIPS 2024 · similarity 0.84)
- IDEAL: Influence-Driven Selective Annotations Empower In-Context Learners in Large Language Models (ICLR 2024 · 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.
