Moritz Zanger

icon company TU Delft icon location Delft icon location My name is Moritz Zanger - PhD Student at the Technical University of Delft Interested in Reinforcement Learning, Uncertainty

Results 2 issues of Moritz Zanger

Clipped Loss in Critic of ppo2 unnecessarily long?

1 comment

Hi, I've been wondering whether the code for the approximate trust region update of the critic (via clipping losses) is a little more convoluted than it has to be. specifically,...

Tensorflow BOOT DQN agent loses performance after first iteration

Hi, I am observing a strange behavior by the tensorflow default boot dqn agent that I am a bit baffled by. When running sweeps over multiple environments, the agent loses...