AI & MLSpecializedAIML-RL-022
Partially Observable MDPs
Acting when the state is hidden: belief states and why memory or filtering becomes part of the policy.
- Difficulty
- 5 / 5
- Importance
- 3 / 5
- Study time
- 1–2 h
- Kind of work
- Mixed
Why it matters
The learner can update a belief state after an observation and explain why an optimal POMDP policy acts on beliefs rather than observations.