Research map / Label-efficient and robust learning / Noisy labels
Linear Distance Metric Learning with Noisy Labels
Meysam Alishahi, Anna Little, Jeff M. Phillips
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 covariance normalization on the combustion data instead of scaling-only normalization (ablation) | Logistic metric learning · Combustion equations-of-state | relative spectral error: 0.748 -> 0.041 | Improved |
| add covariance normalization to the input data (ablation) | Logistic metric learning · Combustion equations-of-state | mean test accuracy: 99.51% -> 99.73% | Improved |
| feed raw unscaled data to the solver (ablation) | Logistic metric learning · Combustion equations-of-state | convergence | Hurt |
| use noise-matched (Logistic) labels instead of direct random label flipping (ablation) | Logistic loss model · Synthetic | test accuracy w/o noise: 98.83% -> 94.51% | Improved |
| use noise model close to Logistic rather than HS noise (ablation) | Logistic loss model · Synthetic | test accuracy w/ noise: 89.85% -> 85.28% | Improved |
| increase mislabeling from 10% to 45% while holding training size (ablation) | Logistic loss model · Synthetic, Logistic noise | accuracy vs ground truth: around 95% at 40% noise; collapses at 45% | Hurt |
| increase training set to 2e5 samples under 45% noise (ablation) | Logistic loss model · Synthetic, 45% Logistic noise | test accuracy vs ground truth: 65% -> 97% | Improved |
| increase number of training points from 100 to about 1000 (ablation) | Logistic loss model · Synthetic | train/test accuracy gap | Improved |
| use the proposed Logistic loss model instead of DML-eig | Metric learning · Synthetic, Logistic noise | test accuracy vs ground truth: 90% -> 99% | Improved |
| use the proposed Logistic loss model instead of DML-eig | Metric learning · Synthetic training time | training time: 1200 seconds -> less than a second | Improved |
| replace Logistic loss with Laplace or HS loss under Logistic noise (ablation) | Metric learning loss · Synthetic eigenvalue recovery | eigenvalue recovery | No clear effect |
Closest papers by meaning
- Efficient Stochastic Optimization for Low-Rank Distance Metric Learning (AAAI 2017 · similarity 0.89)
- Unsupervised Metric Learning with Synthetic Examples (AAAI 2020 · similarity 0.88)
- Learning Low-Dimensional Metrics (NeurIPS 2017 · similarity 0.88)
- Signal-To-Noise Ratio: A Robust Distance Metric for Deep Metric Learning (CVPR 2019 · similarity 0.87)
- Denoising Multi-Similarity Formulation: A Self-Paced Curriculum-Driven Approach for Robust Metric Learning (AAAI 2023 · similarity 0.87)
- Sample-Efficient Geometry Reconstruction from Euclidean Distances using Non-Convex Optimization (NeurIPS 2024 · similarity 0.85)
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.
