CS7642_Week10 Flashcards
1
Q
What is a DEC-POMDP?
A
Decentralized Partially Observable MDP. It’s a way of redefining MDPs that are more suited to coordination.
2
Q
What are three general ways that a human can communicate to an agent?
A
- Demonstrations
- Rewards
- Policies
3
Q
What is a TTD-MDP?
A
Targeted Trajectory MDP
4
Q
TTD-MDPs can be solved in linear time? (True/False)
A
True (linear in the length of the story)