Research map / Label-efficient and robust learning / Recognition and long-tail
Creating Training Sets via Weak Indirect Supervision
Jieyu Zhang, Bohan Wang, Xiangchen Song, Yujing Wang, Yaming Yang, Jing Bai, Alexander Ratner
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 |
|---|---|---|---|
| Model ILF accuracy and fine-grained label-relation strengths explicitly instead of using pseudo-accuracy factors | WS-LG / PLRM · LSHTC-3, label model | Accuracy: 62.60 $\pm$ { 10.12} -> 64.65 $\pm$ { 11.30} | Improved |
| Model ILF accuracy and fine-grained label-relation strengths explicitly instead of using pseudo-accuracy factors | WS-LG / PLRM · LSHTC-3, end model | F1-score: 67.36 $\pm$ { 14.24} -> 69.37 $\pm$ { 14.41} | Improved |
| Model ILF accuracy and fine-grained label-relation strengths explicitly instead of using pseudo-accuracy factors | WS-LG / PLRM · ILSVRC2012, label model | Accuracy: 53.68 $\pm$ { 7.62} -> 56.18 $\pm$ { 7.35} | Improved |
| Model ILF accuracy and fine-grained label-relation strengths explicitly instead of using pseudo-accuracy factors | WS-LG / PLRM · ILSVRC2012, end model | F1-score: 54.57 $\pm$ { 11.17} -> 56.83 $\pm$ { 8.49} | Improved |
| Use PLRM rather than WS-LG to synthesize labels | PLRM · Car Accessories | F1: 66.85 -> 76.37 | Improved |
| Use PLRM rather than WS-LG to synthesize labels | PLRM · Car Accessories | Accuracy: 66.33 -> 75.83 | Improved |
| Use PLRM rather than WS-LG to synthesize labels | PLRM · Furniture | F1: 70.59 -> 80.57 | Improved |
| Use PLRM rather than WS-LG to synthesize labels | PLRM · Furniture | Accuracy: 74.51 -> 82.02 | Improved |
| Violate the distinguishability condition by giving unseen labels identical relations to every seen label | PLRM · WIS tasks with indistinguishable unseen labels | Overall performance: about 10 points | Hurt |
Closest papers by meaning
- Stronger Than You Think: Benchmarking Weak Supervision on Realistic Tasks (NeurIPS 2024 · similarity 0.90)
- Training Complex Models with Multi-Task Weak Supervision (AAAI 2019 · similarity 0.89)
- Characterizing the Impacts of Semi-supervised Learning for Weak Supervision (NeurIPS 2023 · similarity 0.89)
- Learning from weak labelers as constraints (ICLR 2025 · similarity 0.89)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.88)
- Reduction-based Pseudo-label Generation for Instance-dependent Partial Label Learning (NeurIPS 2025 · 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.
