Research map / Learning theory and algorithms / Sample complexity
Randomized Block Cubic Newton Method
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 a random block subset (block size tuned, medium-sized blocks) instead of full-coordinate or very small blocks in Cubic Newton (ablation) | RBCN (Algorithm 1) on synthetic regression · Synthetic regression (N up to 16K) | total computational time to 1e-12 accuracy | Helped |
| Use block size 25-50 instead of larger blocks for Cubic Newton (ablation) | Block coordinate Cubic Newton on logistic regression · leukemia and duke breast-cancer logistic regression | total computational time | Helped |
| Replace block coordinate gradient descent with the cubic Newton block method (with block size 25-50) | Logistic regression with constraints reformulation · leukemia and duke breast-cancer | total computational time | Helped |
| Use primal-dual cubic Newton (Algorithm 2) instead of SDNA and SDCA for Poisson regression | Poisson regression dual ascent · Synthetic (m=1000, d=200) and Montreal bike lanes (m=319, d=20) | epochs to reach given duality gap | Helped |
Closest papers by meaning
- Sub-sampled Newton Methods with Non-uniform Sampling (NeurIPS 2016 · similarity 0.89)
- Stochastic Subspace Cubic Newton Method (ICML 2020 · similarity 0.89)
- Stochastic Variance-Reduced Cubic Regularized Newton Methods (ICML 2018 · similarity 0.88)
- Stochastic Cubic Regularization for Fast Nonconvex Optimization (NeurIPS 2018 · similarity 0.88)
- A Regularized Newton Method for Nonconvex Optimization with Global and Local Complexity Guarantees (NeurIPS 2025 · similarity 0.85)
- The Multi-Block DC Function Class: Theory, Algorithms, and Applications (NeurIPS 2026 · 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.