You are on page 1of 42

Fundamentals of

Artificial Intelligence
KRISHNA MIYAPURAM

Course Topics
What you have Learnt
So far?
Intelligent Agents
Problem Solving Using
Search Techniques

Knowledge Representation

First-order/Predicate Logic
Production Rules
Logic programming (Prolog)
Slot and Filler Structures

(Machine) Learning

Bayesian Networks
Connectionist Networks
Reinforcement Learning

Applications of AI

Natural language processing


Expert Systems
Cognitive Robotics (Perception &
Action)

Resources

Course Policies - Fall 2014


Total Assessment = 50%
Topical seminars (2), Peer teaching = 10%
Class Participation, Lecture summaries , Take-Home

Assignments = 10%
Project (Term paper &/ Presentation) = 10%
Quizzes (surprise / planned) = 10%
End semester Exam = 10%

Lecture Outline
Overview of AI
History
Applications

Artificial Intelligence Tasks


Name a few Artificial Intelligence Tasks

Artificial Intelligence Tasks


Mundane Tasks

Perception: Vision, Speech


Natural Language: Understanding, Generation, Machine Translation
Common Sense Reasoning
Robot Control

Formal Tasks

Games: Chess, Backgammon, Checkers (Go)


Mathematics: Geometry, Logic, Integral calculus, Proving properties
of programs

Expert Tasks

Engineering: Design, Fault finding, Manufacturing planning


Scientific Analysis, Medical Analysis, Financial Analysis

Artificial Intelligence Tasks


General

Learning and Perception

Specific

Playing chess
Proving mathematical theorems
Writing poetry
Driving a car on a crowded street
Diagnosing diseases

What is Artificial Intelligence?


Attempt a definition of Artificial Intelligence

What is Artificial Intelligence?


Artificial Intelligence is the study of how to make

computers do things which, at the moment, people


do better.

Dimensions of Artificial Intelligence


Thought Process and Reasoning

Systems that think rationally


Systems that think like humans

Behaviour

Machines that acts like humans


Machine that acts rationally

A machine that __ is
an intelligent machine

Thought Process and Reasoning


Systems that think rationally

Function by rules
Cannot make mistakes
Logical thinking: Correct premises necessarily lead to correct
conclusions

Interruptions, distractions, emotions

Systems that think like humans

Begin with a theory of human thought process


Express the theory as a computer program
If the programs input output behaviour matches corresponding
human behaviour,
then there is some evidence that some of the programs mechanisms
could also be operating in humans

Behaviour
Thought Process and Reasoning

Systems that think rationally


Systems that think like humans

Behaviour

Rather than endowing machines with the ability to think like humans, get machines
to act like humans, to do what we do

Building a machine that acts rationally

Without necessarily being concerned with whether the machines are doing it in the same
way as humans
An agent is just something that acts
A rational agent is one that acts so as to achieve
the best outcome
under uncertainty, the best expected outcome

It is not necessary to take into account the components of human behaviour (or
human thought)

Notoriously difficult to quantify and recreate in a program

Setting the Stage for Artificial Intelligence


The beginnings
Historical background:

19th Century

The Jaquard Loom


Charles Babbage: difference & analytical engines

20th century

Alan Turing
John von Neumann

Schoty, The Russian Abacus


We use fingers on our

hand and devices such as


abacus to keep track of
numbers and numerical
operations more than
what could be readily
held in our
working memory

The Human Computers


The women of the Computer

Department at NACA HighSpeed Flight Research


Station are busy with test
flight calculations.
The computers were
responsible for accurate
calculations on the research
test flights made at the
Station.
There were no mechanical
computers at the station in
1949.

The Jacquard Loom


Intricate design is weaved onto a fabric

The patterning of the

fabric woven is dictated


by a program that
specifies which threads of
the warp (length-wise)
and the weft (cross-ways)
strands should go over
and under the shuttle.
It made used of punched
cards: a set of cards with
holes in them arranged in
the desired pattern

Charles Babbage (1791-1871)

Alan Turing (1912 1954)

John von Neumann (1903 1957)

The Machine Performs


Early Programs

The Logic Theorist


The General Problem
Solver

Expert Systems

MYCIN
Deep Blue (Chess)

Natural Language

processing

Robotics

The SOAR project

Connectionism

Neural Networks

Semantic Information
processing SHRDLU
Script Applier Mechanism
Speech Recognition
Robot Vision
Animats,, projects Cog,
Domo, Kismet,

Herbert Simon & Alan Newell