Neural temporal-difference learning converges to global optima

INIS

• learning 100%
• neural networks 33%
• convergence 33%
• evaluation 16%
• policy 16%
• optimization 16%
• values 16%
• coupling 16%
• errors 16%
• nonlinear problems 16%
• approximations 16%

Computer Science

• temporal difference learning 100%
• Global Convergence 40%
• Building-Blocks 20%
• Nonlinearities 20%
• Function Approximation 20%
• Deep Reinforcement Learning 20%
• Roles 20%
• Evaluation 20%