Download CS 430 Lecture 5

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Signals intelligence wikipedia , lookup

Artificial intelligence wikipedia , lookup

Transcript
Lecture 5

Reminder: SWE CECS Career Forum sign-up
by today

Reminder: Homework 1 due Monday

Questions?
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
1
Outline

Chapter 2 - Intelligent Agents

Agents and Environments

Concept of Rationality

Nature of Environments
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
2
Agents and Environments

An agent is an entity that can be viewed as
perceiving its environment through sensors,
and acting upon that environment through
actuators.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
3
Agent Function and Agent Program



An agent's perceptual input at any given
instance is called a percept. The complete
history of everything an agent has seen (so far)
is its percept sequence.
An agent's behavior is described by a function
that receives a percept and returns an action.
In theory, this a table that maps all possible
percept sequences to actions.
An agent program is a concrete
implementation of the agent function.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
4
Vacuum-Cleaner World

Percepts


location (square A or square B), cleanliness state
(Clean or Dirty)
Actions

move Left, move Right, Suck dirt
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
5
Concept of Rationality


A rational agent does the "right thing" for every
entry in the agent function table. The sequence
of actions results in environment states that are
desirable, a notion captured by a performance
measure.
Designing performance measures is hard.
Beware of unintended consequences.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
6
Concept of Rationality

Rationality depends on




Performance measure - objective (numeric)
Agent's prior knowledge of environment
Actions that can be performed
Percept sequence to date
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
7
Concept of Rationality

Define rational agent as:




For each possible percept sequence, a rational
agent should select an action that is
expected to maximize its performance measure,
given the evidence provided by the percept
sequence
and whatever built-in knowledge the agent has.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
8
Vacuum-Cleaner World

Does the table we generated represent a
rational agent?



First, we need a performance measure...
Assume the "geography" of environment is known a
priori, but not the distribution of dirt or initial location
of agent. Clean squares stay clean and sucking
cleans a square

Only available actions are Left, Right, Suck

Agent correctly perceives location and dirt.
If any of these "assumptions" are changed?
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
9
Concept of Rationality



Omniscience - knowing the actual outcome of actions
- is not part of rationality. With omniscience, can
achieve perfection and maximize actual performance
vs. expected performance.
Information gathering and exploration is part of
rationality, necessary for initially unknown
environments and to maximize performance.
Agents should learn from perception and become
autonomous - effectively independent of its a priori
knowledge.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
10
Nature of Environments


Task environments are specified by

Performance measure

Environment

Actuators

Sensors
Give a PEAS description for an automated taxi
driver. For a medical diagnosis system. For an
interactive English tutor.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
11
Characteristics of Environments

Fully observable vs. Partially observable


Can the agent "see" the entire environment relevant
to performance measure.
Single agent vs. Multi-agent

Are other participants agents or are they
environment that obey "laws"? Are the other agents
competitive or cooperative? Is there
communication between agents?
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
12
Characteristics of Environments

Deterministic vs. Stochastic


Is the next state of environment completely
determined by the current state and the agent
action? Environments that are not fully observable
or not deterministic are said to be uncertain.
Episodic vs. Sequential

In episodic environments, agent experience is
divided into atomic episodes that to not effect each
other. In sequential environments, current decision
can potentially effect all future decisions.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
13
Characteristics of Environments

Static vs. dynamic


Does the environment change while the agent is
making a decision? If the environment does not
change, but the performance score does, said to be
semidynamic.
Discrete vs. Continuous

Refers to the state of the environment, to the way
time is handled, and to percepts and actions of the
agent.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
14
Characteristics of Environments

Known vs. unknown


Does the agent know the "laws of physics" for the
environment? Note: not the same as fully
observable vs. partially observable.
Not too surprisingly, hardest case is partially
observable, multi-agent, stochastic, sequential,
dynamic, continuous, and unknown
environment. E.g. driving a rented car in a new
country with unfamiliar geography and traffic
laws.
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
15
Characteristics of Environments

What are the characteristics of the
environments for the following?

Solving crossword puzzles

Playing backgammon

Playing poker

Medical diagnosis

Interactive English tutor

Taxi driving
Friday, January 20
CS 430 Artificial Intelligence - Lecture 5
16