Revisit last-iterate convergence of mSGD under milder requirement on step size.
Ruinan JinXingkang HeLang ChenDifei ChengVijay GuptaPublished in: NeurIPS (2022)
Keyphrases
- step size
- convergence rate
- convergence speed
- faster convergence
- variable step size
- line search
- global convergence
- cost function
- learning rate
- global optimum
- adaptive filter
- evolutionary programming
- pso algorithm
- gradient method
- differential evolution
- hessian matrix
- convergence analysis
- stochastic gradient descent
- conjugate gradient
- steepest descent method
- optimization problems
- image classification
- principal component analysis
- primal dual
- particle swarm optimization
- feature space