Research map / Learning theory and algorithms / Matrix and tensor methods
Perceptual adjustment queries and an inverted measurement paradigm for low-rank metric learning
Austin Xu, Andrew McRae, Jingyan Wang, Mark A. Davenport, Ashwin Pananjady
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 PAQ responses instead of pairwise, triplet or ranking-k ordinal queries | low-rank metric learning · simulated d=50, r=10, noiseless | responses needed for error 0.2: 1,000 -> 600 | Improved |
| Increase number of measurements N (ablation) | PAQ estimator · varying d, r=15, y=200, eta=10 | normalized estimation error | Improved |
| Normalize sample size by dimension d (N/d) with fixed rank (ablation) | PAQ estimator · varying d | normalized estimation error | Improved |
| Increase rank r above 8 (ablation) | PAQ estimator · rank sweep | normalized estimation error | Hurt |
| Use rank r <= 8 instead of r > 8 (ablation) | PAQ estimator · rank sweep | normalized estimation error | Hurt |
| Set averaging parameter m proportional to (N/d)^(1/3) (ablation) | PAQ estimator · d=50, r=9, y=200, eta=200 | normalized estimation error | Improved |
| Increase N/d from 400 to 1000 (ablation) | PAQ estimator optimal averaging parameter · d=50, r=9 | optimal averaging parameter m: 400 -> 1000 | Mixed |
Closest papers by meaning
- A Unified Model and Dimension for Interactive Estimation (NeurIPS 2023 · similarity 0.83)
- SDS-LoRA: Overcoming Anisotropic Gradient Scaling in Low-Rank Adaptation (NeurIPS 2026 · similarity 0.82)
- Comparing Comparisons: Informative and Easy Human Feedback with Distinguishability Queries (ICML 2025 · similarity 0.81)
- MAdam: Metric-Aware Multi-Objective Adam (NeurIPS 2026 · similarity 0.81)
- RepLoRA: Reparameterizing Low-rank Adaptation via the Perspective of Mixture of Experts (ICML 2025 · similarity 0.80)
- Item Response Scaling Laws: A Measurement Theory Approach for Efficient and Generalizable Neural Scaling Estimation (ICML 2026 · similarity 0.80)
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.
