CS 365A: Artificial Intelligence
Table of Contents
1 Homework #1 – Warm Up
Submitted by: Karan Singh {11354}
1.1 Cartoon
1.2 Task Identification
1.2.1 Table
Task | Deterministic | Episodic | Static | Discrete | Fully observable | Single agent |
---|---|---|---|---|---|---|
or | or | or | or | or | vs. | |
Stochastic | Sequential | Dynamic | Continuous | Partially observable | Multiagent | |
Playing soccer | Stochastic | Sequential | Dynamic | Continuous | Partially observable | Multiagent |
Brushing your teeth | Deterministic | Sequential | Static | Continuous | Partially observable | Single agent |
Playing a tennis match | Stochastic | Sequential | Dynamic | Continuous | Partially observable | Multiagent |
Practicing tennis againts a wall | Deterministic | Sequential | Dynamic | Continuous | Fully observable | Single agent |
Deciding what item to take at hostel lunch | Deterministic | Episodic | Static | Discrete | Fully observable | Single agent |
1.2.2 Reasons
Let us take up the example of "Playing soccer".
- Stochastic
The task is stochastic since the next state does depend on actions taken by other players too, and is not uniquely determined by one's action.
- Sequential
The action to be taken now depends on past actions. For example, if you kick the ball to the right, you would have to run in the same direction to chase the ball.
- Dynamic
Since there are other players involved and the ball is moving, the environment can change even when the agent is not involved.
- Continuous
Soccer is a continuous-state and continuous-time problem since the speed of the ball and its direction are both continuous quantities.
- Partially observable
All the information that is relevent may not be available to the agent. For example, it is hard to see if someone is tackling you from behind, if you're kicking the ball forward.
- Multiagent
Since the behaviour of the agent depends on how the other players (both team-mates and the opposing team) play, the problem is a multiagent problem.