CS561Lecture2

# CS561Lecture2 - Lecture 2: Intelligent Agents (Part I,...

This preview shows pages 1–13. Sign up to view the full content.

Artifcial Intelligence Lecture 2: Intelligent Agents (Part I, Chapter 2) Spring 2010 Instructor: Paul S. Rosenbloom

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Course Updates New TA Office Sal 229 Phone: (213) 740-5421 Final is May 12, 2-4pm 2
3 Today’s Lecture Agents and environments Introducing the vacuum-cleaner world The concept of rational behavior Environments Agent structure

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
4 Agents and Environments Agents are systems that perceive and act in some environment Include humans, robots, softbots, thermostats, etc. Ignore K & G for now Environment is world in which agent operates Vacuum-Cleaner World
Cognitive Cycle At each point in time, agent must decide what to do next This is the cognitive cycle that repeats as long as the agent lives In humans, the cycle runs at ~50-100ms This is minimum time to choose an action, but many such cycles can be combined to make harder choices On each cycle, agent can be considered to be computing a function 5

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
6 Agent Function The agent function is a mathematical relationship that maps percept sequences to actions in the environment f: P* A For example: [B, Dirty] SUCK [Car 20’ away][Car 10’ away] RUN
7 From Agent Functions to Agent Programs The agent function is computed by an agent program The agent program runs on the physical architecture to implement the function I.e., generates actions from percept sequences

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
8 The Vacuum-Cleaner World Environment: square A and B Percepts: [Location, Content] e.g. [A, Dirty] Actions: left, right, suck, and no-op
A Table-Based Agent Program Percept sequence Action [A,Clean] Right [A, Dirty] Suck [B, Clean] Left [B, Dirty] Suck [A, Clean],[A, Clean] Right [A, Clean],[A, Dirty] Suck

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
10 Return to Rationality What is rational at a given time depends on: Performance measure Ideally objective, external, based on what is to be achieved Prior environment knowledge Actions Percept sequence to date (sensors) DEF: A rational agent chooses whichever action maximizes the expected value of the performance measure given the percept sequence to date and prior environment knowledge
11 What Rationality Isn’t Generality, literacy, autonomy or collaboration Omniscience An omniscient agent knows what will occur Perfection Rationality maximizes expected (prospective) performance, while perfection maximizes actual (retrospective) performance

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
12 Environments Environment: World in which the agent operates To understand agent behavior - or to design a special purpose agent - need to understand its environment Simon’s Ant: Complex behavior may arise from a simple program in a complex environment Rationality defined, at least in part, in terms of agent’s environment PEAS description of the environment: P erformance: Measure for success/progress/quality E nvironment: The world in which the agent operates Environment in the narrow versus the broad context A ctuators: How the agent affects the environment S ensors: How the agent perceives the environment
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 03/05/2010 for the course CS 561 taught by Professor Moradi during the Spring '09 term at USC.

### Page1 / 44

CS561Lecture2 - Lecture 2: Intelligent Agents (Part I,...

This preview shows document pages 1 - 13. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online