Research map / Learning theory and algorithms / Submodular and game algorithms
EigenGame: PCA as a Nash Equilibrium
Ian Gemp, Brian McWilliams, Claire Vernade, Thore Graepel
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 EigenGame instead of Frequent Directions | PCA on small ImageNet (ResNet-50 activations) · Small ImageNet (ResNet-50 activations) | runtime: 9 hours -> 2 hours | Improved |
| Scale EigenGame to the full ResNet-200 activation representation | EigenGame · ImageNet (ResNet-200 activations) | runtime | Mixed |
| Remove the hierarchy (ablation) | EigenGame · Synthetic data | principal components learned | Hurt |
| Remove the hierarchy and normalization (ablation) | EigenGame · Synthetic data | principal components learned | Hurt |
| Use EigenGame rather than Oja's algorithm with QR orthogonalization | PCA on synthetic data · Synthetic data | principal component accuracy | Mixed |
| Use EigenGame with a smaller mini-batch | PCA on MNIST · MNIST training set | longest streak | Hurt |
| Use EigenGame in a high batch-size regime | PCA on MNIST · MNIST training set | longest streak | No clear effect |
| Omit projection of the update direction onto the sphere's tangent space (ablation) | EigenGame update · PCA | stability near equilibrium | Improved |
Closest papers by meaning
- EigenGame Unloaded: When playing games is better than optimizing (ICLR 2022 · similarity 0.87)
- The Symmetric Generalized Eigenvalue Problem as a Nash Equilibrium (ICLR 2023 · similarity 0.86)
- Distributed Principal Component Analysis with Limited Communication (NeurIPS 2021 · similarity 0.86)
- Rethinking PCA Through Duality (NeurIPS 2025 · similarity 0.85)
- Convergence of Stochastic Gradient Descent for PCA (ICML 2016 · similarity 0.85)
- Efficient Sparse PCA via Block-Diagonalization (ICLR 2025 · similarity 0.83)
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.
