AI intro UMK.pptx

7/27/2019 AI intro UMK.pptx

1/47

Artificial Intelligence

BE IT

Urmila Kalshetti


2/47

Objectives

To introduce the basic principles and applicationsof Artificial Intelligence

To Understand the basic areas of artificialintelligence such as problem solving, knowledgerepresentation, reasoning, planning, perception,vision and learning

To develop the ability to design and implementkey components of intelligent agents and expertsystems of moderate complexity

To study different Heuristic Search Techniquesand their applications


3/47

References Text Books:

1. Artificial Intelligence: A Modern Approach by Peter and NorvigISBN-0-13-103805-2,

2. Artificial Intelligence by Elaine Rich, Kevin Knight and Nair ISBN-978-0-07-008770-5, TMH

Reference Books:1. George F. Luger , Artificial Intelligence: Structures and Strategies

for Complex Problem Solving, Pearson, ISBN-10: 0321545893

2. N.P. Padhy, Artificial Intelligence And Intelligent Systems, OxfordUniversity Publishers, ISBN 9780195671544

3. Ivan Bratko, PROLOG : Programming for Artificial Intelligence,Pearson Education, 3 Edition, ISBN10: 0-201-40375-7

4. Saroj Kaushik, Artificial Intelligence, Cengage Learning, , ISBN-13:9788131510995


4/47

Why study AI?

AI helps computer scientists and engineers build more useful

and user-friendly computers,

psychologists, linguists, and philosophers understand

the principles that constitute what we callintelligence.

AI is an interdisciplinary field of study.

Many ideas and techniques now standard in CS

(symbolic computation, time sharing, objects,declarative programming, . . . ) were pioneeredby AI-related research.


5/47

AI is among us!

Recent applications using AI techniques:

Sony Aibo

Entertainment robot with pet-like behaviour

Dragon NaturallySpeaking

(Dictation and voice recognition software)

(http://www.dragonsys.com)

You talk, it types.

to use our voice to create documents, write papers, send email, andsearch the Web

KIROBO Robot project

Japan's Kirobo spacebot performs on video There are already quite a few robots on the International Space

Station (namely, Robonaut and a bunch ofSPHEREs), but later thisyear, a little humanoid from Japan will be joining the team:

TOPIO, a humanoid robot,

played ping pong at

Tokyo International RobotExhibition (IREX) 2009.
http://localhost/var/www/apps/conversion/tmp/scratch_4/Aibo%20-%20Sony%20Robo%20Dog%20%20%20%20-%20YouTube.flvhttp://www.dragonsys.com/http://spectrum.ieee.org/tag/robonauthttp://spectrum.ieee.org/tag/sphereshttp://spectrum.ieee.org/tag/robonauthttp://spectrum.ieee.org/tag/sphereshttp://en.wikipedia.org/wiki/TOPIOhttp://en.wikipedia.org/wiki/Ping_ponghttp://en.wikipedia.org/wiki/International_Robot_Exhibitionhttp://en.wikipedia.org/wiki/International_Robot_Exhibitionhttp://en.wikipedia.org/wiki/International_Robot_Exhibitionhttp://en.wikipedia.org/wiki/International_Robot_Exhibitionhttp://en.wikipedia.org/wiki/Ping_ponghttp://en.wikipedia.org/wiki/TOPIOhttp://spectrum.ieee.org/tag/sphereshttp://spectrum.ieee.org/tag/robonauthttp://www.dragonsys.com/http://localhost/var/www/apps/conversion/tmp/scratch_4/Aibo%20-%20Sony%20Robo%20Dog%20%20%20%20-%20YouTube.flvhttp://localhost/var/www/apps/conversion/tmp/scratch_4/Aibo%20-%20Sony%20Robo%20Dog%20%20%20%20-%20YouTube.flv


6/47

AI is among us! More applications using AI techniques:

Honda Humanoid Robot Demo walking robot

Deep Blue(now retired) a new version (Watson) Over three nights, it took on two of the all-time most successful human

players of the game and beat them in front of millions of television viewers inFEB, 2011.

(http://researchweb.watson.ibm.com/deepblue) Mars Pathfinder (1997)

Autonomous land vehicle sent to Mars

(http://mars.jpl.nasa.gov/MPF)

NASA's Juno Spacecraft Launches to Jupiter

Juno's detailed study of the largest planet in our solar system will helpreveal Jupiter's origin and evolution.

Marvel Real-time expert system for monitoring data sent by Voyager spacecraft.

ChatterBot Eliza She is programmed to behave as a Rogerian psychotherapist, and is an

interesting example of the limitations of early artificial intelligence programs

http://nlp-addiction.com/eliza/
http://mars.jpl.nasa.gov/MPFhttp://nlp-addiction.com/eliza/http://nlp-addiction.com/eliza/http://nlp-addiction.com/eliza/http://nlp-addiction.com/eliza/http://nlp-addiction.com/eliza/http://nlp-addiction.com/eliza/http://nlp-addiction.com/eliza/http://nlp-addiction.com/eliza/http://mars.jpl.nasa.gov/MPF


7/47

What is AI?

A scientific and engineering discipline devoted

to:

Understanding principles that make intelligent

behavior possible in natural or artificial systems;

Developing methods for the design and

implementation of useful, intelligent artifacts

Artificial Intelligence is concerned with thedesign of intelligence in an artificial device.


8/47

What is AI?

What is intelligence?

Is it that which characterize humans? Or is there anabsolute standard of judgement?

Accordingly there are two possibilities:

A system with intelligence is expected to behave as intelligentlyas a human

A system with intelligence is expected to behave in the bestpossible manner

Secondly what type of behavior are we talking about?

Are we looking at the thought process or reasoning ability of thesystem?

Or are we only interested in the final manifestations of thesystem in terms of its actions?


9/47

What is AI?

different interpretations have been used by

different researchers

AI is about designing systems that are as

intelligent as humans.

understand human thought and an effort to build

machines that emulate the human thought process

The concept of the Turing Test


10/47

Turing Test


11/47

AI is pretty hard stuff!

I went to the grocery store, I saw the milk on theshelf and I bought it.

What did I buy? The milk?

The shelf?

The store?

An awful lot of knowledge of the world is needed toanswer simple questions like this one.


12/47

Typical AI problems

Recognizing people, objects.

Communicating (through natural language).

Navigating around obstacles on the streets Expert tasks include:

Medical diagnosis.

Mathematical problem solving Playing games like chess


13/47

What is AI?

Views of AI fall into four categories:

Thinking humanly Thinking rationallyActing humanly Acting rationally

The textbook advocates "acting rationally"Rationally means sensibly, logically.


14/47

Acting humanly: Turing Test Turing (1950) "Computing machinery and intelligence":

"Can machines think?" "Can machines behaveintelligently?"

Operational test for intelligent behavior: the Imitation Game

Predicted that by 2000, a machine might have a 30% chance

of fooling a lay person for 5 minutes Anticipated all major arguments against AI in following 50

years

Suggested major components of AI: knowledge, reasoning,language understanding, learning


15/47

Thinking humanly: cognitive modeling

1960s "cognitive revolution": information-processingpsychology

Getting inside the actual working of human minds Introspection

Pshychological experiments

-- How to validate? Requires1) Predicting and testing behavior of human subjects (top-down)

or 2) Direct identification from neurological data (bottom-up)

Both approaches (roughly, Cognitive Science andCognitive Neuroscience) are now distinct from AI


16/47

Thinking rationally: "laws of thought"

Aristotle: what are correct arguments/thought processes?(right thinking)

Socrates is a man. All men are mortal Therefore Socrates is mortal.

Several Greek schools developed various forms oflogic:notation and rules of derivation for thoughts.

Direct line through mathematics and philosophy to modernAI

Challenges:1. Stating informal knowledge in formal in formal terms

2. Solving a problem in principle and doing so in practice


17/47

Acting rationally: rational agent

Rational behavior: doing the right thing

The right thing: that which is expected tomaximize goal achievement, given the availableinformation

Doesn't necessarily involve thinking e.g.,blinking reflex but thinking should be in theservice of rational action

Laws of thought- emphasis is on correct inference

Correct inference is not all of rationality becausethere are often situations where there is noprovably correct thing to do, yet something muststill be done.


18/47

AI prehistory

Philosophy Logic, methods of reasoning, mind as physicalsystem foundations of learning, language,rationality

Mathematics Formal representation and proof algorithms,computation, (un)decidability, (in)tractability,probability

Economics utility, decision theory

Neuroscience physical substrate for mental activity

Psychology phenomena of perception and motor control,experimental techniques

Computer building fast computers

engineering Control theory design systems that maximize an objective

function over time

Linguistics knowledge representation, grammar


19/47

Abridged history of AI

1943 McCulloch & Pitts: Boolean circuit model of brain

1950 Turing's "Computing Machinery and Intelligence"

1956 Dartmouth meeting: "Artificial Intelligence" adopted

195269 Look, Ma, no hands!

1950s Early AI programs, including Samuel's checkers

program, Newell & Simon's Logic Theorist,Gelernter's Geometry Engine

1965 Robinson's complete algorithm for logical reasoning

196673 AI discovers computational complexityNeural network research almost disappears

196979 Early development of knowledge-based systems

1980-- AI becomes an industry 1986-- Neural networks return to popularity

1987-- AI becomes a science

1995-- The emergence of intelligent agents


20/47

Task Domains of AI

Routine Tasks Perception

Natural language Understanding, generation, translation

Commonsense reasoning

Robot control

Formal Tasks Games

Chess, Backgammon, Checkers

Mathematics Geometry, logic, integral calculus

Expert Tasks Engineering

Designing, fault finding, manufacturing planning

Scientific analysis

Financial analysis

Medical Analysis


21/47

Agents

An agent is anything that can be viewed as perceivingits environment through sensors and acting uponthat environment through actuators

Human agent: eyes, ears, and other organs forsensors; hands, legs, mouth, and other body partsfor actuators

Robotic agent: cameras and infrared range finders forsensors; various motors for actuators


22/47

Examples of agents

Humans can be looked upon as agents. Theyhave eyes, ears, skin, taste buds, etc. forsensors; and hands, fingers, legs, mouth for

effectors .

Robots are agents. Robots may have camera,

sonar, infrared, bumper, etc. for sensors. Theycan have grippers, wheels, lights, speakers,etc. for actuators.


23/47

Agents and environments

The agentfunction maps from percept histories to actions:

[f: P*A]

The agentprogram runs on the physical architecture to

producef agent = architecture + program


24/47

Vacuum-cleaner world

Percepts: location and contents, e.g., [A,Dirty]

Actions: Left, Right, Suck, NoOp


25/47

A vacuum-cleaner agent

Percept Sequence The complete history of everything the agent has ever

perceived

Agent function Maps any given percept sequence to an action

Abstract mathematical description

Agent program Concrete implementation running on the agent

architecture


26/47

Vacuum-cleaner world

Simple agent function

What is the right way to fill out the table?

What makes an agent good or bad, intelligent or stupid?

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

.

.[A, clean], [A, clean], [A, clean] Right

[A, clean], [A, clean], [A, Dirty] Suck


27/47

Rational agents

An agent should strive to "do the right thing", based on whatit can perceive and the actions it can perform. The right actionis the one that will cause the agent to be most successful.

Performance measure: An objective criterion for success of anagent's behavior

E.g., performance measure of a vacuum-cleaner agent could

be amount of dirt cleaned up, amount of time taken, amountof electricity consumed, amount of noise generated, etc.


28/47

Rational agents

RationalAgent: For each possible percept

sequence, a rational agent should select an

action that is expected to maximize its

performance measure, given the evidenceprovided by the percept sequence and

whatever built-in knowledge the agent has.


29/47

Rational agents

Rationality is distinct from omniscience (all-knowing with infinite knowledge)

Agents can perform actions in order to modifyfuture percepts so as to obtain useful information(information gathering, exploration)

An agent is autonomous if its behavior isdetermined by its own experience (with ability tolearn and adapt)


30/47

PEAS

PEAS: Performance measure, Environment,Actuators, Sensors

Consider, e.g., the task of designing an automatedtaxi driver:

Performance measure

Environment

Actuators Sensors


31/47

PEAS

Consider, e.g., the task of designing an automatedtaxi driver:

Performance measure: Safe, fast, legal, comfortable trip,

maximize profits Environment: Roads, other traffic, pedestrians, customers

Actuators: Steering wheel, accelerator, brake, signal, horn

Sensors: Cameras, sonar, speedometer, GPS, odometer,engine sensors, keyboard


32/47

PEAS

Agent: Medical diagnosis system

Performance measure: Healthy patient,

minimize costs, lawsuits

Environment: Patient, hospital, staff

Actuators: Screen display (questions, tests,

diagnoses, treatments, referrals)

Sensors: Keyboard (entry of symptoms,

findings, patient's answers)


33/47

PEAS

Agent: Part-picking robot

Performance measure: Percentage of parts in

correct bins

Environment: Conveyor belt with parts, bins

Actuators: Jointed arm and hand

Sensors: Camera, joint angle sensors


34/47

PEAS

Agent: Interactive English tutor

Performance measure: Maximize student's

score on test

Environment: Set of students

Actuators: Screen display (exercises,

suggestions, corrections)

Sensors: Keyboard


35/47

Software agents/Softbots Softbot designed to fly a flight simulator for a large

commercial airplane.

Softbot designed to scan internet news sources andshow the interesting items to its customers It will need NLP ability

It will need to learn what each customer is interested in It will need change its plan dyanmically

soft robots based on natural forms, including squidand starfish. Whitesides envisions using the

pneumatically powered robots to aid disaster recoveryefforts by squeezing into the rubble left by anearthquake to locate survivors, or as a way to free up asurgeons hands in the operating room.http://news.harvard.edu/gazette/story/2011/12/soft-

bots/
http://news.harvard.edu/gazette/story/2011/12/soft-bots/http://news.harvard.edu/gazette/story/2011/12/soft-bots/http://news.harvard.edu/gazette/story/2011/12/soft-bots/http://news.harvard.edu/gazette/story/2011/12/soft-bots/http://news.harvard.edu/gazette/story/2011/12/soft-bots/


36/47

Environment types

Fully observable (vs. partially observable): An agent's sensors give itaccess to the complete state of the environment at each point intime. Partially observable- if noisy and inaccurate sensors

Deterministic (vs. stochastic): The next state of the environment iscompletely determined by the current state and the actionexecuted by the agent. (If the environment is deterministic exceptfor the actions of other agents, then the environment is strategic)

Episodic (vs. sequential): The agent's experience is divided intoatomic "episodes" (each episode consists of the agent perceivingand then performing a single action), and the choice of action ineach episode depends only on the episode itself.


37/47

Environment types

Static (vs. dynamic): The environment is unchangedwhile an agent is deliberating. (The environment issemidynamic if the environment itself does notchange with the passage of time but the agent's

performance score does)

Discrete (vs. continuous): A limited number ofdistinct, clearly defined percepts and actions.

Single agent (vs. multiagent): An agent operating byitself in an environment.


38/47

Environment types

Chess with Chess without Taxi drivinga clock a clock

Fully observable Yes Yes No

Deterministic Strategic Strategic No

Episodic No No No

Static Semi Yes NoDiscrete Yes Yes No

Single agent No No No

The environment type largely determines the agent design

The real world is (of course) partially observable, stochastic, sequential,

dynamic, continuous, multi-agent


39/47

Agent types

Four basic types in order of increasing

generality:

Simple reflex agents

Model-based reflex agents

Goal-based agents

Utility-based agents


40/47



41/47


\input{algorithms/d-agent-algorithm}


42/47



43/47


\input{algorithms/d+-agent-algorithm}


44/47

Goal-based agents


45/47

Utility-based agents


46/47

Learning agents


47/47

Chatterbot Eliza

Documents

AI intro UMK.pptx