Research map / Learning theory and algorithms / Sample complexity
Robust Unsupervised Learning via L-statistic Minimization
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 |
|---|---|---|---|
| Use the robust L-statistic objective instead of standard k-means | K-means on synthetic data with outliers · Synthetic 2D three-cluster data | center recovery | Helped |
| Use the robust L-statistic objective instead of standard k-means | K-means with misspecified k · Synthetic 2D three-cluster data, k=2 | center recovery | Helped |
| Use the robust L-statistic objective instead of standard k-means | K-means on Fashion-MNIST · Fashion-MNIST sneakers and trousers with other classes as outliers | reconstruction error | Helped |
| Use the robust L-statistic objective instead of standard k-means | K-means on Fashion-MNIST · Fashion-MNIST sneakers and trousers with other classes as outliers | center quality | Helped |
| Use the robust L-statistic objective instead of standard principal subspace analysis | Principal subspace analysis on synthetic data · Synthetic 2D data with outliers | subspace recovery | Helped |
| Use the robust L-statistic objective instead of standard principal subspace analysis | Principal subspace analysis on Fashion-MNIST · Fashion-MNIST sneakers and trousers with other classes as outliers | reconstruction error | Helped |
| Vary the hard-threshold cutoff ζ across the tested range (ablation) | RKM on Fashion-MNIST · Fashion-MNIST sneakers and trousers with other classes as outliers | reconstruction error | No clear effect |
| Vary the hard-threshold cutoff ζ across the tested range (ablation) | RPSA on Fashion-MNIST · Fashion-MNIST sneakers and trousers with other classes as outliers | reconstruction error | No clear effect |
| Use the robust L-statistic objective instead of standard k-means | K-means on clean data · Clean data | performance | No clear effect |
Closest papers by meaning
- Better generalization with less data using robust gradient descent (ICML 2019 · similarity 0.87)
- Tight Risk Bounds for Gradient Descent on Separable Data (NeurIPS 2023 · similarity 0.87)
- Algorithmic Stability Unleashed: Generalization Bounds with Unbounded Losses (ICML 2024 · similarity 0.86)
- Multiclass Loss Geometry Matters for Generalization of Gradient Descent in Separable Classification (NeurIPS 2025 · similarity 0.85)
- Out-Of-Domain Unlabeled Data Improves Generalization (ICLR 2024 · similarity 0.85)
- On the Provable Advantage of Unsupervised Pretraining (ICLR 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.