Research map / Probabilistic and scientific modeling / Bayesian optimization
Hybrid-MST: A Hybrid Active Sampling Strategy for Pairwise Preference Aggregation
JING LI, Rafal Mantiuk, Junle Wang, Suiyi Ling, Patrick Le Callet
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 |
|---|---|---|---|
| switch from global-maximum EIG sampling to MST sampling after one standard trial | Hybrid-MST · Simulated data and real-world VQA/IQA | rating aggregation (PLCC and RMSE) | Improved |
| use Hodge-active instead of Hybrid-MST for limited-budget ranking aggregation | pairwise preference aggregation · Simulated data and real-world VQA/IQA, 2-3 standard trial numbers | Kendall | Improved |
| use Hybrid-MST rather than Crowd-BT | pairwise preference aggregation · Real-world VQA and IQA | PLCC and RMSE | Improved |
| use Hybrid-MST rather than Hodge-active | pairwise preference aggregation · Real-world VQA and IQA, increasing test budget | Kendall | Mixed |
| use MST mode rather than GM mode | Hybrid-MST · Simulated data, n=10 | runtime (ms/pair): 48.72 -> 6.16 | Improved |
| use MST mode rather than GM mode | Hybrid-MST · Simulated data, n=20 | runtime (ms/pair): 153.61 -> 8.97 | Improved |
| use MST mode rather than GM mode | Hybrid-MST · Simulated data, n=100 | runtime (ms/pair): 3007.08 -> 30.04 | Improved |
| use batch-mode MST sampling rather than sequential sampling | Hybrid-MST · Typical VQA pair comparison, ideal parallel execution | time for n-1 pairs: 15.1; 15.2; 17.9 | Improved |
| use Hodge-active rather than Hybrid-MST for small-budget ranking aggregation | pairwise preference aggregation · Simulated dataset with small test budget | Kendall | Improved |
Closest papers by meaning
- Active preference learning for ordering items in- and out-of-sample (NeurIPS 2024 · similarity 0.84)
- Adaptive Hardness Negative Sampling for Collaborative Filtering (AAAI 2024 · similarity 0.83)
- Leveraging Sparsity for Sample-Efficient Preference Learning: A Theoretical Perspective (ICML 2025 · similarity 0.83)
- Preference learning made easy: Everything should be understood through win rate (ICML 2025 · similarity 0.83)
- Robust Sampling for Active Statistical Inference (NeurIPS 2025 · similarity 0.82)
- Bridging and Modeling Correlations in Pairwise Data for Direct Preference Optimization (ICLR 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.
