Research map / Optimization / Optimizers
Towards Practical Alternating Least-Squares for CCA
Zhiqiang Xu, Ping Li
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 |
|---|---|---|---|
| couple the two least-squares problems sequentially with block size k (TALS) instead of CCALin | CCALin · Mediamill, JW11, MNIST | sin^2 theta_u and sin^2 theta_v vs time and passes | Improved |
| use block size k instead of 2k for plain ALS and CCALin (ALS-k, CCALin-k) (ablation) | ALS and CCALin · Mediamill, JW11, MNIST | sin^2 theta_u and sin^2 theta_v | Hurt |
| add momentum term with beta set after T0 burn-in TALS iterations (FALS-T0) | TALS · Mediamill, JW11, MNIST | sin^2 theta_u and sin^2 theta_v | Improved |
| adaptively set momentum parameter during iterations (AALS) instead of burn-in tuning | FALS-T0 · Mediamill, JW11, MNIST | sin^2 theta_u and sin^2 theta_v | No clear effect |
| burn-in T0 = 6 versus T0 = 4 for FALS (ablation) | FALS · Mediamill, JW11, MNIST | sin^2 theta_u and sin^2 theta_v | No clear effect |
| add faster ALS to vector case versus shift-and-invert preconditioning | FALS · vector case k=1 (supplementary) | convergence | Mixed |
Closest papers by meaning
- Efficient Globally Convergent Stochastic Optimization for Canonical Correlation Analysis (NeurIPS 2016 · similarity 0.87)
- Efficient Algorithms for Large-scale Generalized Eigenvector Computation and Canonical Correlation Analysis (ICML 2016 · similarity 0.85)
- Stochastic Approximation for Canonical Correlation Analysis (NeurIPS 2017 · similarity 0.85)
- Stochastic Canonical Correlation Analysis (JMLR 2019 · similarity 0.84)
- Unconstrained Stochastic CCA: Unifying Multiview and Self-Supervised Learning (ICLR 2024 · similarity 0.83)
- On Sparse Canonical Correlation Analysis (NeurIPS 2024 · similarity 0.82)
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.
