Adaptive SGD with Polyak stepsize and Line-search: Robust Convergence and Variance Reduction.
Xiaowen JiangSebastian U. StichPublished in: NeurIPS (2023)
Keyphrases
- line search
- step size
- convergence rate
- variance reduction
- convergence speed
- cost function
- faster convergence
- global convergence
- estimation error
- global optimum
- machine learning
- primal dual
- sample size
- objective function
- monte carlo
- particle swarm optimization
- quadratic programming
- importance sampling
- multiresolution
- conjugate gradient