Research map / Label-efficient and robust learning / Recognition and long-tail
A Universal Unbiased Method for Classification from Aggregate Observations
Zixi Wei, Lei Feng, Bo Han, Tongliang Liu, Gang Niu, Xiaofeng Zhu, Heng Tao Shen
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 UUM instead of log-likelihood | CFAO with pairwise similarity · MNIST | test accuracy: 98.87\pm 0.06\%\bullet -> 98.99 \pm 0.04\% | Improved |
| Use UUM instead of log-likelihood | CFAO with triplet comparison · CIFAR-10 | test accuracy: 70.57 \pm 0.59\%\bullet -> 72.41 \pm 0.98\% | Improved |
| Use UUM instead of ROT | Learning from label proportions · Fashion-MNIST | test accuracy: 89.40 \pm 0.31\%\bullet -> 89.87 \pm 0.20\% | Improved |
| Use UUM instead of log-likelihood | Multiple-instance learning · Musk2 | group-level test accuracy: 80.20\pm 12.70\%\bullet -> 99.61\pm 0.37\% | Improved |
| Use UUM instead of log-likelihood | CFAO with pairwise similarity · vehicle | test accuracy: 79.41 \pm 4.23 \% -> 78.71 \pm 4.20 \% | Hurt |
| Remove log-likelihood warm-up initialization (ablation) | UUM for pairwise similarity · CIFAR-10 | test accuracy: 71.45 \pm 0.47\% -> 73.45\pm 0.40\% | Improved |
Closest papers by meaning
- Bipartite Ranking From Multiple Labels: On Loss Versus Label Aggregation (ICML 2025 · similarity 0.84)
- Learning with Selectively Labeled Data from Multiple Decision-makers (ICML 2025 · similarity 0.84)
- Learning from Aggregate responses: Instance Level versus Bag Level Loss Functions (ICLR 2024 · similarity 0.83)
- Learning from Label Proportions: Bootstrapping Supervised Learners via Belief Propagation (ICLR 2024 · similarity 0.83)
- Class-Attribute Priors: Adapting Optimization to Heterogeneity and Fairness Objective (AAAI 2024 · similarity 0.83)
- PriorBoost: An Adaptive Algorithm for Learning from Aggregate Responses (ICML 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.
