3 ms·
Some more discussion on the difficulty of creating reward functions: https://www.alexirpan.com/2018/02/14/rl-hard.html#reinforcement-learning-usually-requires-a
by wetmore 6y ago
Some more discussion on the difficulty of creating reward functions: https://www.alexirpan.com/2018/02/14/rl-hard.html#reinforcement-learning-usually-requires-a-reward-function https://www.alexirpan.com/2018/02/14/rl-hard.html#reinforcem...
- _0ffh 6y agoLol, that link sent me on a nearly one hour detour and then sent me back to HN (specifically, to [1])! https://news.ycombinator.com/item?id=6269114 https://news.ycombinator.com/item?id=6269114