Enrique Mallada

gradient descent (2) linear network (2) reinforcement learning (1) spectral clustering (1) graph laplacian (1) global convergence (1) policy learning (1) restricted isometry property (1) constrained markov decision process (1) convergence rate (1) weight initialization (1) gradient flow (1) implicit bia (1) model reduction (1) exponential convergence (1) stationary policy (1) dynamic network (1) subspace-preserving recovery (1) multi-layer linear model (1) gradient dominance property (1)