A Review of Partially Observable Markov Decision Processes for Causal Modeling

author: Finale Doshi-Velez, Harvard Medical School
published: Oct. 6, 2014,   recorded: December 2013,   views: 1978

See Also:

Download slides icon Download slides: nipsworkshops2013_doshi_velez_modeling_01.pdf (269.6┬áKB)

Help icon Streaming Video Help

Related Open Educational Resources

Related content

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Lecture popularity: You need to login to cast your vote.


Partially Observable Markov Decison Processes (POMDPs) are a framework for modeling sequential decision-making problems. At every time-step, an agent takes an action that causes some (hidden) state of the world to change. The hidden state then emits some observations and rewards that the agent may use to guide its next decision. In this way, POMDPs provide a useful model for actions that may have long-range temporal effects. For example, when entering a building, an initial decision to turn left or right will place an agent in very different places, even if all the agent's future actions are to continue forward.

Link this page

Would you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !

Write your own review or comment:

make sure you have javascript enabled or clear this field: