Research map / Learning theory and algorithms / Matrix and tensor methods
Dropping Symmetry for Fast Symmetric Nonnegative 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 |
|---|---|---|---|
| Use SymANLS instead of the strongest listed alternative | Symmetric NMF clustering · ORL | clustering accuracy: 0.7700 -> 0.8075 | Helped |
| Use SymANLS instead of the strongest listed alternative | Symmetric NMF clustering · COIL-20 | clustering accuracy: 0.7472 -> 0.7979 | Helped |
| Use SymHALS instead of the strongest listed alternative | Symmetric NMF clustering · MNIST_train | clustering accuracy: 0.6477 -> 0.6657 | Helped |
| Use SymHALS instead of the strongest listed alternative | Symmetric NMF clustering · TDT2 | clustering accuracy: 0.9803 -> 0.9806 | Helped |
| Use SymGCD instead of SymANLS | Symmetric NMF clustering · MNIST_test | clustering accuracy: 0.8589 -> 0.9882 | Hurt |
| Tune the regularization parameter λ; very large or very small values slow convergence (ablation) | SymHALS · Synthetic data, n = 50, r = 5 | convergence speed: 39.9 | Mixed |
| Use the theoretically sufficient λ bound as a necessary threshold (ablation) | SymHALS · Synthetic data, n = 50, r = 5 | factor agreement: 39.9 | Hurt |
| Transfer symmetric NMF to a regularized two-factor nonsymmetric formulation and use alternating updates | Symmetric NMF · Synthetic data and CBCL | convergence speed | Helped |
| Use alternating methods rather than SymNewton | Symmetric NMF image clustering · ORL and COIL-20 | running time | Helped |
| Use ADMM rather than an alternating method | Symmetric NMF image clustering · ORL | iterations to comparable clustering accuracy: 3500 | Hurt |
Closest papers by meaning
- An Exterior Method for Nonnegative Matrix Factorization (ICML 2026 · similarity 0.89)
- Non-negative Matrix Factorization under Heavy Noise (ICML 2016 · similarity 0.89)
- Statistically Optimal $K$-means Clustering via Nonnegative Low-rank Semidefinite Programming (ICLR 2024 · similarity 0.88)
- Characterizing the Loss Landscape in Non-Negative Matrix Factorization (AAAI 2021 · similarity 0.88)
- Local Centroids Structured Non-Negative Matrix Factorization (AAAI 2017 · similarity 0.88)
- Low Rank Matrix Completion via Robust Alternating Minimization in Nearly Linear Time (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.