Projected Stochastic Gradient Langevin Algorithms for Constrained Sampling and Non-Convex Learning

Keyphrases

• Additive Noise 25%
• Constrained Sampling 100%
• Constraint Set 25%
• Convergence Property 25%
• Convex Compact Set 25%
• Convex Constraints 25%
• External Data 25%
• Gradient Method 25%
• Langevin Algorithm 100%
• Learning Problems 25%
• Log-concave Distributions 50%
• Monte Carlo (MC) Simulation 25%
• Monte Carlo Learning 25%
• Monte Carlo Optimization 25%
• Non-convex Cost Function 50%
• Non-convex Learning 100%
• Nonconvex Optimization 25%
• Sample Learning 100%
• Sampling Optimization 25%
• Stochastic Gradient 100%
• Suboptimal Solutions 25%
• Super-exponential 25%
• Target Distribution 25%
• Wasserstein Distance 25%

Mathematics

• Additive Noise 50%
• Compact Set 50%
• Convergence Property 50%
• Markov Chain Monte Carlo 50%
• Polynomial 50%
• Stochastics 100%
• target distribution π 50%
• Wasserstein Distance 50%