Re:Deep Learning in Robotics: A Review of Recent Research
by hackpert on Tuesday July 25, 2017 @03:50PM
(#14851275)
> Anyone else going to the DeepRL Bootcamp at Berkeley in August ?
I was really looking forward to it, especially because among other things, I wanted to discuss the very arguments of the absence of true randomness in simulations that you mentioned below in context of RL -- it seems there hasn't been a lot of work in analyzing the situation with simulations lately because everyone tends to get excited upon seeing the algorithm figure out bipedal movement in a variety of simulated environments et cetera. Out of sheer bad luck, I have something else scheduled around the same dates that I cannot skip :(. Oh well.
I was really looking forward to it, especially because among other things, I wanted to discuss the very arguments of the absence of true randomness in simulations that you mentioned below in context of RL -- it seems there hasn't been a lot of work in analyzing the situation with simulations lately because everyone tends to get excited upon seeing the algorithm figure out bipedal movement in a variety of simulated environments et cetera. Out of sheer bad luck, I have something else scheduled around the same dates that I cannot skip :(. Oh well.