A New Supermemory Gradient Method without Line Search for Unconstrained Optimization.
June LiuHuanbin LiuYue ZhengPublished in: ISNN (4) (2009)
Keyphrases
- unconstrained optimization
- gradient method
- line search
- step size
- convergence rate
- trust region
- convergence speed
- global convergence
- conjugate gradient
- cost function
- primal dual
- learning rate
- faster convergence
- risk minimization
- wavelet coefficients
- knn
- feature space
- objective function
- global optimum
- neural network
- recommender systems
- pairwise