MathDB1 problem
- 0 votes0 replies0 views
Conjecture on the convergence time of continuous-action reinforcement learning
Convergence-time conjecture. The convergence time, when the algorithm converges, is very high even for a quadratic cost function.
Conjecture on the convergence time of continuous-action reinforcement learning
Convergence-time conjecture. The convergence time, when the algorithm converges, is very high even for a quadratic cost function.