AI and Agents

BAD402
Module 1

1
What is Artificial Intelligence?
 Definitions of AI vary
 Artificial Intelligence is the study of
systems that
think like humans think rationally

act like humans act rationally

2
Systems Acting like Humans
 Turing test: test for intelligent behavior
 Interrogator writes questions and receives
answers
 System providing the answers passes the test if
interrogator cannot tell whether the answers come
from a person or not
 Necessary components of such a system form
major AI sub-disciplines:
 Natural language, knowledge representation,
automated reasoning, machine learning

3
Systems Thinking like Humans
 Formulate a theory of mind/brain
 Express the theory in a computer
program
 Two Approaches
 Cognitive Science and Psychology (testing/
predicting responses of human subjects)
 Cognitive Neuroscience (observing
neurological data)
4
Systems Thinking Rationally
 “Rational” -> ideal intelligence
(contrast with human intelligence)
 Rational thinking governed by precise
“laws of thought”
 syllogisms
 notation and logic
 Systems (in theory) can solve problems
using such laws

5
Systems Acting Rationally
 Building systems that carry out actions
to achieve the best outcome
 Rational behavior
 May or may not involve rational thinking
 i.e., consider reflex actions

 This is the definition we will adopt

6
Intelligent Agents
 Agent: anything that perceives and
acts on its environment
 AI: study of rational agents
 A rational agent carries out an action
with the best outcome after
considering past and current percepts

7
Foundations of AI
 Philosophy: logic, mind, knowledge
 Mathematics: proof, computability, probability
 Economics: maximizing payoffs
 Neuroscience: brain and neurons
 Psychology: thought, perception, action
 Control Theory: stable feedback systems
 Linguistics: knowledge representation, syntax

8
Brief History of AI
 1943: McCulloch & Pitts: Boolean circuit
model of brain
 1950: Turing's “Computing Machinery and
Intelligence”
 1952—69: Look, Ma, no hands!
 1950s: Early AI programs, including Samuel's
checkers program, Newell & Simon's Logic
Theorist, Gelernter's Geometry Engine
 1956: Dartmouth meeting: “Artificial
Intelligence” adopted
9
Brief History of AI
 1965: Robinson's complete algorithm for
logical reasoning
 1966—74: AI discovers computational
complexity; Neural network research almost
disappears
 1969—79: Early development of knowledge-
based systems
 1980—88: Expert systems industry booms
 1988—93: Expert systems industry busts:
`”AI Winter”
10
Brief History of AI
 1985—95: Neural networks return to
popularity
 1988— Resurgence of probability;
general increase in technical depth,
“Nouvelle AI”: ALife, GAs, soft
computing
 1995— Agents…

11
Back to Agents

12
Agent Function
 a = F(p)
where p is the current percept, a is the action
carried out, and F is the agent function
 F maps percepts to actions
F: P  A
where P is the set of all percepts, and A is the set of
all actions
 In general, an action may depend on all
percepts observed so far, not just the current
percept, so…

13
Agent Function Refined
 ak = F(p0 p1 p2 …pk)
where p0 p1 p2 …pk is the sequence of
percepts observed to date, ak is the
resulting action carried out
 F now maps percept sequences to
actions
F: P*  A

14
Structure of Agents
 Agent = architecture + program
 architecture
 device with sensors and actuators
 e.g., A robotic car, a camera, a PC, …
 program
 implements the agent function on the
architecture

15
Specifying the Task
Environment
 PEAS
 Performance Measure: captures agent’s
aspiration
 Environment: context, restrictions
 Actuators: indicates what the agent can
carry out
 Sensors: indicates what the agent can
perceive
16
Properties of Environments
 Fully versus partially observable
 Deterministic versus stochastic
 Episodic versus sequential
 Static versus dynamic
 Discrete versus continuous
 Single agent versus multiagent

17
Example: Mini Casino world
 Two slot machines
 Costs 1 peso to play in a machine
 Takes 10 seconds to play in a machine
 Possible pay-offs: 0, 1, 5, 100
 Given:
 Amount of money to start with
 Amount of time to play
 Expected payoff for each machine
 Objective: end up with as much money as
possible
Mini Casino World
 PEAS description?
 Properties
 Fully or partially observable?
 Deterministic or stochastic?
 Episodic or sequential?
 Static or dynamic?
 Discrete or continuous?
 Single agent or multi-agent?
Types of Agents
 Reflex Agent
 Reflex Agent with State
 Goal-based Agent
 Utility-Based Agent

 Learning Agent

20
Reflex Agent

21
Reflex Agent with State

22
State Management
 Reflex agent with state
 Incorporates a model of the world
 Current state of its world depends on
percept history
 Rule to be applied next depends on
resulting state
 state’  next-state( state, percept )
action  select-action( state’, rules )
23
Goal-based Agent

24
Incorporating Goals
 Rules and “foresight”
 Essentially, the agent’s rule set is
determined by its goals
 Requires knowledge of future
consequences given possible actions
 Can also be viewed as an agent with
more complex state management
 Goals provide for a more sophisticated
next-state function

25
Utility-based Agent

26
Incorporating Performance
 May have multiple action sequences
that arrive at a goal
 Choose action that provides the best
level of “happiness” for the agent
 Utility function maps states to a
measure
 May include tradeoffs
 May incorporate likelihood measures

27
Learning Agent

28
Incorporating Learning
 Can be applied to any of the previous
agent types
 Agent <-> Performance Element
 Learning Element
 Causes improvements on agent/
performance element
 Uses feedback from critic
 Provides goals to problem generator

29
Next: Problem Solving
Agents (Chap 3-6)