Research map / Learning theory and algorithms / Prediction and decision losses
Pairwise HITS: Quality Estimation from Pairwise Comparisons in Creator-Evaluator Crowdsourcing Process
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 |
|---|---|---|---|
| add creator ability modeling (Two-Stage vs Pairwise HITS) (ablation) | Pairwise HITS · Image description | Spearman correlation, 1 evaluator per pair: 0.800 ± 0.029 -> 0.857 ± 0.020 | Helped |
| add creator ability modeling (Two-Stage vs Pairwise HITS) (ablation) | Pairwise HITS · Logo designing | Spearman correlation, 1 evaluator per pair: 0.735 ± 0.028 -> 0.766 ± 0.023 | Helped |
| add creator ability modeling (Two-Stage vs Pairwise HITS) (ablation) | Pairwise HITS · Language translation | Spearman correlation, 1 evaluator per pair: 0.604 ± 0.047 -> 0.661 ± 0.035 | Helped |
| add creator ability modeling (Two-Stage vs Pairwise HITS) (ablation) | Pairwise HITS · Language translation | Spearman correlation, 5 evaluators per pair: 0.885 ± 0.019 -> 0.877 ± 0.017 | Hurt |
| use Pairwise HITS instead of Bradley-Terry | Bradley-Terry model · Image description | Spearman correlation, 1 evaluator per pair: 0.782 ± 0.031 -> 0.800 ± 0.029 | Helped |
| use Pairwise HITS instead of Bradley-Terry | Bradley-Terry model · Logo designing | Spearman correlation, 5 evaluators per pair: 0.924 ± 0.009 -> 0.927 ± 0.009 | No clear effect |
| use Crowd-BT instead of Bradley-Terry | Bradley-Terry model · Logo designing | Spearman correlation, 5 evaluators per pair: 0.924 ± 0.009 -> 0.838 ± 0.019 | Hurt |
| use Crowd-BT instead of Bradley-Terry | Bradley-Terry model · Image description | Spearman correlation, 1 evaluator per pair: 0.782 ± 0.031 -> †0.796 ± 0.029 | Helped |
Closest papers by meaning
- A Minimax Optimal Algorithm for Crowdsourcing (NeurIPS 2017 · similarity 0.85)
- Avoiding Imposters and Delinquents: Adversarial Crowdsourcing and Peer Prediction (NeurIPS 2016 · similarity 0.85)
- Strategyproof Peer Selection: Mechanisms, Analyses, and Experiments (AAAI 2016 · similarity 0.85)
- Efficient Online Crowdsourcing with Complex Annotations (AAAI 2024 · similarity 0.83)
- Label Aggregation for Composite Crowd Tasks by Worker Ability Constraint Satisfaction (AAAI 2025 · similarity 0.82)
- Evaluating LLM-contaminated Crowdsourcing Data Without Ground Truth (NeurIPS 2025 · similarity 0.82)
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.