Research map / Optimization / Implicit bias and generalization
Characterizing the Loss Landscape in Non-Negative Matrix Factorization
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 |
|---|---|---|---|
| Increase ResNet width from k=1 to k=4 (ablation) | ResNet-32 · CIFAR10, epoch 300 | convexity length scale: 0.55 ˘ 0.037 -> 0.82 ˘ 0.032 | Helped |
| Increase ResNet width from k=1 to k=4 (ablation) | ResNet-44 · CIFAR10, epoch 300 | convexity length scale: 0.57 ˘ 0.036 -> 0.78 ˘ 0.032 | Helped |
| Increase ResNet width from k=1 to k=4 (ablation) | ResNet-56 · CIFAR10, epoch 300 | convexity length scale: 0.57 ˘ 0.035 -> 0.79 ˘ 0.033 | Helped |
| Increase ResNet width from k=1 to k=4 (ablation) | ResNet-68 · CIFAR10, epoch 300 | convexity length scale: 0.58 ˘ 0.036 -> 0.81 ˘ 0.033 | Helped |
| Increase ResNet width from k=1 to k=4 (ablation) | ResNet-80 · CIFAR10, epoch 300 | convexity length scale: 0.63 ˘ 0.036 -> 0.79 ˘ 0.03 | Helped |
| Increase ResNet width from k=1 to k=4 (ablation) | ResNet-110 · CIFAR10, epoch 300 | convexity length scale: 0.63 ˘ 0.037 -> 0.82 ˘ 0.03 | Helped |
| Increase the sampled NMF matrix size by using more rows (ablation) | NMF · Birdsong, MovieLens, satellite, ORL faces | fraction of sampled curvatures that are positive | No clear effect |
| Increase the sampled NMF matrix size by using more rows (ablation) | NMF · Birdsong, MovieLens, satellite, ORL faces | relative deviation of curvature | Helped |
Closest papers by meaning
- Supervised Matrix Factorization: Local Landscape Analysis and Applications (ICML 2024 · similarity 0.87)
- In-depth Analysis of Low-rank Matrix Factorisation in a Federated Setting (AAAI 2025 · similarity 0.87)
- Provable Acceleration of Nesterov's Accelerated Gradient for Asymmetric Matrix Factorization and Linear Neural Networks (NeurIPS 2024 · similarity 0.87)
- Creating Coherence in Federated Non-Negative Matrix Factorization (AAAI 2025 · similarity 0.87)
- On the Crucial Role of Initialization for Matrix Factorization (ICLR 2025 · similarity 0.87)
- Statistically Optimal $K$-means Clustering via Nonnegative Low-rank Semidefinite Programming (ICLR 2024 · similarity 0.86)
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.