Research map / Learning theory and algorithms / Matrix and tensor methods
CoLA: Exploiting Compositional Structure for Automatic and Efficient Numerical Linear Algebra
Andres Potapczynski, Marc Anton Finzi, Geoff Pleiss, Andrew Gordon Wilson
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 |
|---|---|---|---|
| split product structure into a sequence of solves via a product composition rule | CoLA iterative solver · Bi-Poisson PDE | runtime | Improved |
| use SVRG (VR-PCA) instead of SVD or Lanczos for large N | CoLA PCA · PCA on Buzz (437.4K, 77) | runtime | Improved |
| exploit sum structure with SVRG for the neural PDE linear system | CoLA · Neural PDEs | training efficiency | Improved |
| use doubly stochastic diagonal estimator instead of standard Hutchinson | CoLA diagonal estimation · Buzz empirical covariance diagonal | relative error vs passes | Improved |
| exploit Kronecker, block diagonal and concatenation structure for equivariant layer construction | CoLA nullspace computation · Equivariant neural network construction | runtime | Improved |
| apply CoLA to applications without compositional structure versus specialized libraries | CoLA · PCA, regression, GP, spectral clustering, Schrodinger, minimal surface | runtime | Mixed |
Closest papers by meaning
- Searching for Efficient Linear Layers over a Continuous Space of Structured Matrices (NeurIPS 2024 · similarity 0.86)
- RL4RLA: Teaching ML to Discover Randomized Linear Algebra Algorithms Through Curriculum Design and Graph-Based Search (ICML 2026 · similarity 0.85)
- Learning to Discover Iterative Spectral Algorithms (NeurIPS 2026 · similarity 0.84)
- Composing Linear Layers from Irreducibles (NeurIPS 2025 · similarity 0.83)
- The Multi-Block DC Function Class: Theory, Algorithms, and Applications (NeurIPS 2026 · similarity 0.83)
- CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs (NeurIPS 2026 · 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.
