Research map / Learning theory and algorithms / Kernels and regression theory
Algorithmic stability and generalization of an unsupervised feature selection algorithm
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 |W_I| instead of W_I^2 as feature selector (ablation) | Ours · MNIST-Fashion | classification accuracy (%): 80.8 -> 80.5 | No clear effect |
| use |W_I| instead of W_I^2 as feature selector (ablation) | Ours · Mice Protein | classification accuracy (%): 99.1 -> 97.2 | Hurt |
| use |W_I| instead of W_I^2 as feature selector (ablation) | Ours · GLIOMA | classification accuracy (%): 80.0 -> 90.0 | Helped |
| use |W_I| instead of W_I^2 as feature selector (ablation) | Ours · Average over 10 datasets | classification accuracy (%): 84.6 -> 87.9 | Helped |
| increase training sample size n | Ours · MNIST-Fashion | test error | Helped |
| increase training sample size n | Ours · MNIST-Fashion | error difference (generalization gap) | Helped |
| increase regularization lambda_1 | Ours · MNIST-Fashion | error difference | Helped |
| increase lambda_1 from 0 to about 0.01 | Ours · MNIST-Fashion | test error | Mixed |
| increase number of selected features k | Ours · MNIST-Fashion | error difference | Hurt |
| increase number of selected features k | Ours · MNIST-Fashion | test error | Helped |
| reduce original features by more than 80% using |W_I| | Ours · 10 datasets | classification accuracy: less than 10% | Hurt |
| increase training set size from 3,000 to 6,000 | Ours · MNIST-Fashion | uniform stability bound | Helped |
Closest papers by meaning
- SAND: One-Shot Feature Selection with Additive Noise Distortion (ICML 2025 · similarity 0.87)
- Building a stable classifier with the inflated argmax (NeurIPS 2024 · similarity 0.86)
- Balancing Feature Similarity and Label Variability for Optimal Size-Aware One-shot Subset Selection (ICML 2024 · similarity 0.84)
- Probabilistic Stability Guarantees for Feature Attributions (NeurIPS 2025 · similarity 0.84)
- Hypernetworks for Dynamic Feature Selection (NeurIPS 2026 · similarity 0.84)
- Most Influential Subset Selection: Challenges, Promises, and Beyond (NeurIPS 2024 · similarity 0.84)
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.