Large Stepsize Gradient Descent for Logistic Loss: Non-Monotonicity of the Loss Improves Optimization Efficiency.
Jingfeng WuPeter L. BartlettMatus TelgarskyBin YuPublished in: CoRR (2024)
Keyphrases
- step size
- cost function
- stochastic gradient descent
- faster convergence
- quasi newton
- steepest descent method
- approximate dynamic programming
- optimization problems
- global optimization
- optimization algorithm
- objective function
- optimization method
- optimization methods
- convergence speed
- convergence rate
- global optimum
- estimation error
- temporal difference
- global convergence
- search direction
- wavelet transform