Hacker News new | ask | show | jobs
by Sommer 2340 days ago
If it did I would have learned from my mistakes.
3 comments

He could have overfit due to his training size being too small. Either he has to add some noise or increase his set.
Maybe your reward function is just wonky.