 Hi everyone, this is Alice Gao. In this short video, I'm going to talk about two variations of a Markov decision process, a fully observable MDP and a partially observable MDP. The two variations, one is a fully observable MDP also just called MDP. You can use the short name versus a partially observable MDP. And this is also called a POMDP. The difference between the two variation is whether each state is observable by the agent. We already encountered a similar issue when we talked about a hidden Markov model. Right in the hidden Markov model, each state is not directly observable, but the agent can observe some noisy signal of each state. So in a fully observable MDP, we assume each state is fully observable. In other words, the agent knows what state it is currently in. In a POMDP, on the other hand, the agent does not know what the state it is currently in, but it can get some noisy signal of the current state. Intuitively, you can think about a POMDP as combining an MDP and a hidden Markov model. In this course, I'll focus on talking about an MDP since it's easier to solve. In practice, a POMDP is much more useful. There are many more scenarios where the state is not directly observable and we should use a POMDP to model it. That's everything for this video. After watching this video, you should be able to explain what's the difference between an MDP and a POMDP. Thank you very much for watching. I will see you in the next video. Bye for now.