Research map / Learning theory and algorithms / Kernels and regression theory
Random matrices in service of ML footprint: ternary random features with no performance loss
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 |
|---|---|---|---|
| Replace RFF with moment-matched TRF at 90% projection-matrix sparsity (ablation) | Random-features ridge regression · MNIST digits 7,9 | test MSE: 90% | No clear effect |
| Replace LP-RFF with TRF | Logistic regression · Cov-Type | test accuracy: 8× less memory | Helped |
| Replace Nyström Gaussian-kernel approximation with TRF | Logistic regression · Cov-Type | test accuracy: 32× or 16× less memory | Helped |
| Replace LP-RFF and Nyström features with TRF | Ridge regression · Census | test MSE | Helped |
| Use ternary weights and ternary activations instead of full-precision random features | Random-features computation and storage · Random-features pipelines | memory: b=32 | Helped |
| Use sparse ternary projection weights with sparsity ε | TRF feature computation · Random-features pipelines | additions required: O((1-ε) mnp) | Helped |
Closest papers by meaning
- A random matrix analysis of random Fourier features: beyond the Gaussian kernel, a precise phase transition, and the corresponding double descent (NeurIPS 2020 · similarity 0.90)
- Simplex Random Features (ICML 2023 · similarity 0.88)
- Orthogonal Random Features (NeurIPS 2016 · similarity 0.87)
- Manifold Random Features (NeurIPS 2026 · similarity 0.86)
- Variance-Reducing Couplings for Random Features (ICLR 2025 · similarity 0.85)
- General Graph Random Features (ICLR 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.