Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Hyper Label Model for Programmatic Weak Supervision
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 |
|---|---|---|---|
| replace data-generation method with prior LF-generation method (ablation) (ablation) | HLM training data · 14 datasets unsupervised | accuracy/F1 average | Helped |
| replace GNN architecture with MLP/DeepSet (best of two) (ablation) | HLM architecture · 14 datasets unsupervised | accuracy/F1 average | Helped |
| replace better-than-random assumption with per-class each-LF-better-than-random assumption (ablation) | HLM assumption · 14 datasets unsupervised | accuracy/F1 average | Helped |
| HLM vs best baseline CLL: single forward pass without dataset-specific learning | label aggregation · 14 datasets unsupervised | accuracy/F1 average: +1.4 points over CLL | Helped |
| HLM vs CLL: single forward pass without dataset-specific learning | label aggregation · 14 datasets unsupervised | running time: about 6x faster than fastest baseline on average | Helped |
| fine-tune pretrained HLM on small set of ground-truth labels (semi-supervised) | HLM · semi-supervised, N_gt>40 | accuracy | Helped |
| HLM applied to unseen datasets with dataset-specific tuning absent | HLM on Census/Youtube · Census and YouTube | F1/accuracy | Helped |
Closest papers by meaning
- Statistical Analysis of an Adversarial Bayesian Weak Supervision Method (NeurIPS 2025 · similarity 0.91)
- Universalizing Weak Supervision (ICLR 2022 · similarity 0.91)
- Losses over Labels: Weakly Supervised Learning via Direct Loss Construction (AAAI 2023 · similarity 0.89)
- Fusing Conditional Submodular GAN and Programmatic Weak Supervision (AAAI 2024 · similarity 0.88)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.87)
- Weak Supervision Performance Evaluation via Partial Identification (NeurIPS 2024 · similarity 0.87)
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.