Incomplete Information Dynamic Games
Incomplete Information

Partially Observable Markov Decision Process (POMDP)
- S - State space
- T:S×A×S→R - Transition probability distribution
- A - Action space
- R:S×A→R - Reward
- O - Observation space
- Z:S×A×S×O→R - Observation probability distribution
Alleatory
Epistemic (Static)
Epistemic (Dynamic)
Partially Observable Markov Game
Alleatory
Epistemic (Static)
Epistemic (Dynamic)
Interaction
- S - State space
- T(s′∣s,a) - Transition probability distribution
- Ai,i∈1..k - Action spaces
- Ri(s,a) - Reward function
- Oi,i∈1..k - Observation space
- Z(oi∣a,s′) - Observation probability distribution
Hierarchy of Problems

Belief updates?
Reduction to Simple Game
Pruning in Dynamic Programming
Extensive Form Game
(Alternative to POMGs that is more common in the literature)
- Similar to a minimax tree for a turn-taking game
- Chance nodes
- Information sets
King-Ace Poker Example
- 4 Cards: 2 Aces, 2 Kings
- Each player is dealt a card
- P1 can either raise (r) the payoff to 2 points or check (k) the payoff at 1 point
- If P1 raises, P2 can either call (c) Player 1's bet, or fold (f) the payoff back to 1 point
- The highest card wins
Extensive to Matrix Form


Exponential in number of info states!
A more interesting example: Kuhn Poker

Fictitious Play in Extensive Form Games

This slide not covered on exam

Heinrich et al. 2015 "Fictitious Self Play in Extensive-Form Games"
Deep Stack: Scaling to Heads Up No Limit Texas Hold 'Em

Can game learning methods like CFR be used in Large POMGs?

Title Text
Next Year: Add papers like the stratego paper, AlphaStar, etc.
Incomplete Information Dynamic Games
255 Incomplete Information Games
By Zachary Sunberg
255 Incomplete Information Games
- 243