Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability — mentioned by YeGoblynQueenne

Thu, 05 Jun 2025 21:14:10 GMT

Probably not. RL is really bad at generalising to unseen environments. There was a paper about an ... otter? Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability https://arxiv.org/abs/2107.06277 OK, it's a robotic zookeeper looking for the otter cage. Where does it say they had no prior knowledge before the event? I can't find that in the text. Is it in the video? I guess there's no paper yet.

Paper and discussion history. A mention is not an endorsement.

Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability — mentioned by YeGoblynQueenne

Mon, 27 May 2024 08:41:01 GMT

>> "If you can build a program which follows a set of rules, and the rules for that language can be differentiated, could you not code a simulation in that differentiable language and then identify the optimal policy using it's gradients?" What's a "policy" here? In optimal control (and reinforcement learning) a policy is a function from a set of states to a set of actions, each action a transition between states. In a program synthesis context I guess that translates to a function from a set of _program_ states to a set of operations? What is an "optimal" policy then? One that transitions…

Paper and discussion history. A mention is not an endorsement.

Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability · HackerPapers discussions

Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability — mentioned by YeGoblynQueenne

Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability — mentioned by YeGoblynQueenne