Transcript Agent.pps

EA C461 – Artificial Intelligence
Intelligent Agents
S.P.Vimal
http://discovery.bits-pilani.ac.in/~vimalsp/1910AI/
To discuss…
• Agents
• Rational Agents
• Task Environments
Vimal
EA C461- Artificial Intelligence
Agent
“An agent is anything that can be viewed
as perceiving its environment through
sensors and acting upon that environment
through actuators”
“An agent’s choice of action at any given
instant can depend on the entire percept
sequence observed to date”
Percept sequence  Complete history of
everything the agent have ever perceived
Vimal
EA C461- Artificial Intelligence
Agent
• An agents behavior is described by agent
function, mapping of any given percept sequence
to an action
• An agent program internally implements the
artificial agent’s agent function
Vimal
EA C461- Artificial Intelligence
A Simple Intelligent Agent
Percept Sequence
• Percepts: location
and contents, e.g.,
[A, Dirty]
• Actions: Left,
Right, Suck, NoOp
Vimal
Action
[A, Clean]
Right
[A, Dirty]
Suck
[B, Clean]
Left
[B, Dirty]
Suck
[A, Clean], [A, Clean]
Right
[A, Clean], [A, Dirty]
Suck
…
[A, Clean], [A, Clean] ,[A,
Clean]
Right
[A, Clean], [A, Clean] , [A,
Dirty]
Suck
…
Simple Agent Function
EA C461- Artificial Intelligence
Good behavior (being Rational)
• Rational (http://dictionary.reference.com)
– agreeable to reason; reasonable; sensible: a rational
plan for economic development.
– having or exercising reason, sound judgment, or good
sense: a calm and rational negotiator.
– being in or characterized by full possession of one's
reason; sane; lucid: The patient appeared perfectly
rational.
– endowed with the faculty of reason: rational beings.
– of, pertaining to, or constituting reasoning powers: the
rational faculty.
– proceeding or derived from reason or based on
reasoning: a rational explanation.
– …
Vimal
EA C461- Artificial Intelligence
Good behavior (being Rational)
• A rational agent is the one that does the
right thing
– Right action is the one that cause the agent to
be more successful
• Appropriate performance measures for the
vacuum world
– Amount of dirt cleaned in a duration
– Clean squares at ever time slots
• How do we say an action is rational at a
given point of time?
Vimal
EA C461- Artificial Intelligence
Rational Agents
For each possible percept sequence, a
rational agent should select an action that is
expected to maximize it’s performance
measure ,given the evidence provided by
the percept sequence and whatever the
built-in knowledge, the agent has.
– Is our Vacuum Cleaner Agent rational?
Vimal
EA C461- Artificial Intelligence
Rational Agents
• Is an omniscient agent rational?
– Maximizing actual performance Vs. Maximizing
expected performance
Learning  autonomy
• A rational agent is autonomous
– Learn to compensate for partial / incorrect
knowledge
Vimal
EA C461- Artificial Intelligence
Task Environment
• Task environments are “problems” for
which the rational agents are the
“solutions”
• Includes
Performance measure
Environment
Actuator
Sensors
Vimal
EA C461- Artificial Intelligence
Task Environment
Agent
Type
Performance
Measures
Environme Actuators
nt
Sensors
Taxi
Driver
Safe, Fast,
Legal, Comfort,
Maximize
Profits
Roads,
other
traffic,
pedestrians
, customers
Steering,
accelerators
, brake,
signal, horn
Camera,
sonar, GPS,
Speedomete
r, keyboard,
etc
Screen
display
(questions,
tests,
diagnoses,
treatments,
referrals)
Keyboard
(entry of
symptoms,
findings,
patient's
answers)
Medical Healthy patient, Patient,
diagnosi minimize costs, hospital,
s
lawsuits
staff
system
PEAS Descriptions
Vimal
EA C461- Artificial Intelligence
Properties of Task Environment
• Fully Observable (vs. Partly Observable)
– Agent sensors give complete state of the environment at
each point in time
– Sensors detect all the aspect that are relevant to the
choice of action
• Deterministic (vs. Stochastic)
– Next state of the environment is completely determined
by the current state and the action executed by the
agent
– Strategic environment
Vimal
EA C461- Artificial Intelligence
Properties of Task Environment
• Episodic (vs. Sequential)
– Agent’s experience can be divided into episodes, each
episode with what an agent perceive and what is the
action
• Next episode does not depend on the previous episode
– Current decision will affect all future sates in sequential
environment
• Static (vs. Dynamic)
– Environment doesn’t change as the agent is deliberating
– Semi dynamic
Vimal
EA C461- Artificial Intelligence
Properties of Task Environment
• Discrete (vs. Continuous)
– Depends the way time is handled in describing state,
percept, actions
• Chess game : discrete
• Taxi driving : continuous
• Single Agent (vs. Multi Agent)
– Competitive, cooperative multi-agent environments
– Communication is a key issue in multi agent
environments
Vimal
EA C461- Artificial Intelligence
Task Environment
Chess with a
clock
Chess without
a clock
Taxi Driving
Fully observable
Deterministic
Episodic
Static
Discrete
Single agent
Example of Task Environments and Their Classes
Vimal
EA C461- Artificial Intelligence
Structure of Agents
Agent = Architecture
+
Computing device Running
Agent Program, with
sensors & actuators
Vimal
Program
Implements Agent Function,
performs mapping of
percepts to actions
EA C461- Artificial Intelligence
Vimal
EA C461- Artificial Intelligence