![Vacuum-cleaner world
• Percepts:
Location and status,
e.g., [A,Dirty]
• Actions:
Left, Right, Suck, NoOp
Example vacuum agent program:
function Vacuum-Agent([location,status]) returns an action
• if status = Dirty then return Suck
• else if location = A then return Right
• else if location = B then return Left](https://image.slidesharecdn.com/ai-lec2-agents-230129180954-989154c9/85/AI-Lec2-Agents-pptx-8-320.jpg)
AI-Lec2-Agents.pptx
- 1. Agents
- 2. Intelligent Agents • Rational agent: one that behaves as well as possible • This behavior depends on the environment • Some environments are more difficult than others
- 3. Agents and Environments • An agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators • An agents behavior is described by the agent function that maps any given percept sequence to an action
- 5. Rational Agents • In this course we will focus on Rational Agents An agent is just something that acts (agent comes from the Latin agere, to do). A rational agent is one that acts so as to achieve the best outcome or, when there is uncertainty, the best expected outcome
- 6. How to describe an Agent • What is the Environment? • What type of Sensors it requires? • Which Actuators are required? • What Percepts it is getting via sensors from environment? – Percept Sequence • Agent Function (map percepts or percept sequence to action)? – Agent Program • Performance Measure: that evaluated the effect of actions
- 7. Example of Agent • Agent: Vacuum Cleaner • Environment: Area A and B • Sensor: Camera • Percept: Area clean or not • Actuator: • Action: Move left, Move Right, – Start cleaning • Agent Function: on next slide • Performance Measure?
- 8. Vacuum-cleaner world • Percepts: Location and status, e.g., [A,Dirty] • Actions: Left, Right, Suck, NoOp Example vacuum agent program: function Vacuum-Agent([location,status]) returns an action • if status = Dirty then return Suck • else if location = A then return Right • else if location = B then return Left
- 10. Example of Agent • Agent: Email Spam filter • Environment: Inbox • Sensor: • Percept: Email • Actuator: • Action: Move email to spam or inbox • Agent Function: Classification Model • Performance Measure: Accuracy, Precision, Recall
- 11. Agent: Spam filter • Performance measure – Minimizing false positives, false negatives • Environment – A user’s email account • Actuators – Mark as spam, delete, etc. • Sensors – Incoming messages, other information about user’s account
- 14. Properties of Environment • Fully observable vs. Partially observable • Deterministic vs. stochastic • Episodic vs. Sequential • Static vs. Dynamic • Discrete vs. Continuous • Single vs. Multivalent
- 15. Types of Envitonment • Reading Assignment • Section 2.3 (Stuart Russell)
- 16. • hardest case is partially observable, multiagent, stochastic, sequential, dynamic, continuous, and unknown
- 17. Four kinds of Agents • Simple Reflex Agent – act only on current percept. • Model Based Reflex Agent. – How the world works. Percept sequence. • Goal based Agent – Act to fulfill some goal. • Utility agent – Act to maximize a utility function. • Learning Agent – Learn from environment and feed back on actions
- 18. Notes for further reading
Editor's Notes
- #11: Precision and recall In pattern recognition, information retrieval and binary classification, precision (also called positive predictive value) is the fraction of relevant instances among the retrieved instances, while recall (also known as sensitivity) is the fraction of relevant instances that have been retrieved over the total amount of relevant instances. Both precision and recall are therefore based on an understanding and measure of relevance.
- #15: Fully observable vs. Partially observable: Complete state of environment is observable Deterministic vs. stochastic: next state is completely determined by the current action Episodic vs. Sequential: Agents experience is divided into atomic episodes, agent perceives and performs single action based on only current state. Static vs. Dynamic: If environment changes while agent is deliberating (semi-dynamic when environment doesn’t change but score of agent changes with passage of time) Discrete vs. Continuous Single vs. Multivalent