Artificial Intelligence

1
Artificial Intelligence

• Lecture 3

2
Topics
Intelligent Agents
I
S
L
P
R
M
N
R
F
P
P
P
P
3
Last Lecture

• Agent and Environment
• Rationality
• World Description

4
Overview

• Review
• Agent Types
• Environment Types

5
Agent Environment
6
Rational Agent

• Agent
• An entity that perceives and acts
• A function from percept histories to actions
• f P -gt A
• Human, robots, thermostat, smoke detector, etc.
• Rational Agent
• For any given set of environments and actions, we
  seek the agent (or class of agent) with the best
  performance

7
Vacuum-Cleaner World

• Percepts location and status, e.g., A,Dirty
• Actions Left, Right, Suck, NoOp

8
Vacuum-Cleaner Agent

• Percept sequence
• Everything the agent has observed so far
• Possible percept sequences
• Mapping percept to action

function Vacuum-Cleaner-Agent(L,S) returns an
action If Sdirty, then return suck Else if LA,
then return right Else if LB, then return left
9
Rationality
10
Performance Measure

• The criteria that determine how successful an
  agent is
• Imposed by authority and measured in the long run

11
Performance Measure

• 100 points for each piece of dirt vacuumed up
• Minus 1 point for each action taken
• Minus 1000 points for dumping the dirt in your
  neighbors backyard

A rational agent maximizes the points given the
percept sequence Rational ? Omniscient Rational ?
Clairvoyant Rational ? Successful Rational gt
Exploration, learning, autonomy
12
World Description
13
Task Environment

• To design a rational taxi agent
• Performance measure
• safety, destination, profits, legality, comfort,
  
• Environment
• US streets/freeways, traffic, pedestrians,
  weather,
• Actuators
• steering, accelerator, brake, horn,
  speaker/display,
• Sensors
• video, accelerometers, gauges, engine sensors,
  keyboard, GPS,

14
Rational Online Auction Agent

• Performance measure?
• Environment ?
• Actuators?
• Sensors?

15
Agent Types
16
Agent Types

• Four basic types in order of increasing
  generality
• Simple reflex agents
• reflex agents with state
• goal-based agents
• utility-based agents
• All these can be turned into learning agents

17
Simple Reflex Agent

• Condition Action Rule
• If condition then action

18
Simple Reflex Function
function Simple-Reflex-Agent(percept) returns
action static rules, a set of condition-action
rules state ? Interpret-Input(percept) rule ?
Rule-Match(state, rules) action ?
Rule-Actionrule return action
19
Reflex Agents with State

• Keep track of the world

20
Reflex Function with Internal States
function Reflex-Agent-With-State(percept) returns
action static state, a description of the
current world state rules, a set of
condition-action rules state ?
Update-State(state, percept) rule ?
Rule-Match(state, rules) action ?
Rule-Actionrule state ? Update-State(state,
action) return action
21
Goal-Based Agents
22
Utility-Based Agents

• Utility a function that maps a state to a real
  value, which describes the degree of
  satisfaction.

23
Learning Agents

24
A Learning Taxi Driver

• Performance element
• Knowledge and procedures for driving
• Learning element
• Formulate goals like learning geography and how
  to drive on wet roads
• Critic
• Observes the world and passes information to the
  learning element
• Problem generator
• Try a different route to see if it is quicker

25
Environment Types
26
Environment Types

• Accessible vs. inaccessible
• Sensory input provides access to the complete
  state of the environment or not
• Accessible environment no need for an internal
  state
• Deterministic vs. nondeterministic
• Next state of the environment is completely
  defined by the current state and the actions of
  the agent
• No uncertainty in an accessible, deterministic
  environment
• An inaccessible environment may appear
  nondeterministic
• Episodic vs. non-episodic
• Each episode consists of an agent perceiving and
  then acting
• Subsequent episodes do not depend on the actions
  in previous episodes no need to think ahead
• A chess tournament is episodic moves in one game
  dont carry over

27
Environment Types

• Static vs. dynamic
• Environment can change while the agent is
  deliberating or not
• In static environments, the agent doesnt need to
  keep looking
• Semi-dynamic environment doesnt change, but the
  agents performance score does
• Discrete vs. continuous
• Limited number of distinct, clearly defined
  percepts and actions
• Chess is discrete as there is a fixed number of
  moves at each turn
• The most difficult type of environment
• Inaccessible, non-deterministic, non-episodic,
  dynamic, continuous

28
Environment Types
29
Environment Types

• The environment type largely determines the
  agent design
• The real world is (of course) partially
  observable, stochastic, sequential, dynamic,
  continuous, multi-agent

30
The Environment Program
procedure Run-Environment(state, Update-Fn,
agents, termination) inputs state, the initial
state of the environment Update-Fn,
function to modify the environment agents,
a set of agents termination, a predicate to
test when we are done repeat for each agent in
agents do Perceptagent ?
Get-Percept(agent, state) end for each agent
in agents do Actionagent ?
Programagent(Perceptagent) end state ?
Update-Fn(actions, agents, state) until
termination(state)
31
Performance of Agents
function Run-Eval-Environment(state, Update-Fn,
agents, termination, Performance-Fn) returns
scores local variables scores, a vector the
same size as agents, all 0 repeat for each
agent in agents do Perceptagent ?
Get-Percept(agent, state) end for each agent
in agents do Actionagent ?
Programagent(Perceptagent) end state ?
Update-Fn(actions, agents, state) scores ?
Performance-Fn(scores, agents, state) until
termination(state) return scores
32
Environment Classes

• Environment class
• The agent must work in different environments
• A chess program should play against a wide
  collection of humans and other programs
• Designing for a particular opponent can exploit
  specific weaknesses, but is not good for general
  play
• The performance of an agent is averaged over the
  environment class
• The agent is not allowed to consult the
  environment program.

33
Summary

• Agent and Environment
• Rationality
• World Description
• Agent Types
• Environment Types

34
Possible Quiz Questions

• If there is a quiz next time, it might cover
• Define rationality
• Difference between reflex agents and goal-based
  utility-based agents
• Components in a learning agent
• Recognize the environment type
• Relation between environment and agents