• 4 Posts
  • 1.04K Comments
Joined 3 years ago
cake
Cake day: June 4th, 2023

help-circle












  • The main difficulty is in how many hyperparameters are involved in training an RL agent, high sensitivity of RL algorithms to those hyperparameters, and not having a good understanding of how to select them based on the properties of your task. This problem is exacerbated by the high sample complexity of RL. If something doesn’t work out, you don’t know if it’s because you chose the wrong set of hyperparameters or if you just haven’t trained for long enough.

    I don’t know much about game design, but I do know that it’s a much more mature field than RL, so surely they have better tools than guessing and praying.







  • The only thing in this list that I relate to is the masking, so I’ll address that

    Yes, we all conform and hide parts of ourselves in public, doesn’t mean you can’t “be yourself”.

    As I understand it, when a neurotypical person is hiding parts of themselves, it’s something like “I don’t want anyone to know I’m into Taylor Swift”. So just don’t talk about it. It’s that simple. For me, I have to think about every word I say because no one interprets things literally. If someone asks me whether or not I like Taylor Swift and I want to answer in the affirmative, can I just say yes? Or do I need to take note of the day of the week and say no while gesturing wildly with my left hand when it’s a Monday or look 15º to the left from Wednesday to Friday? When we talk about masking, it’s that, applied to every single sentence coming out of your mouth. Comparatively, never talking about Taylor Swift is a trivial task.