MODULE 1

 AI problems are classified

 Area of Mathematics

 Commonsense reasoning

 Perception

 Natural Language Understanding

Module 1 AI Problems 2
Area of Mathematics
 To prove Mathematical theorems using games

 Geometry

 Game playing and theorem proving share the property that the

people who do them well are considered to be displaying

intelligence.

Module 1 AI Problems 3
Commonsense Reasoning

 Includes

 Reasoning physical objects and their relationships to

each other

 An object can be in only one place at a time

 Reasoning about actions and their consequences

 If you let go of something, it will fall to the floor and may be

break

Module 1 AI Problems 4
 To investigate this sort of reasoning, NeWell, Shaw and

Simon developed the General Problem Solver (GPS), which

they applied to solve several commonsense tasks as well as

to the problem of performing symbolic manipulations of

logical expressions.

 Selected simple tasks

Module 1 AI Problems 5
Perception

 Perception around our world is crucial our survival.

 Perceptual tasks are difficult because they involve analog

signals rather than digital signals.

 The analog signals are typically very noisy and usually a

large number of things must be perceived at once.

Module 1 AI Problems 6
Natural Language Understanding

 The ability to use language to communicate a wide variety

of ideas is perhaps the most important thing.

 The problem of understanding spoken language is a

perceptual problem and is hard to solve for some reasons.

 We can simplify the problem by restricting it to written

language.

Module 1 AI Problems 7
 This problem, natural language understanding, is also

difficult.

 In order to understand sentences about the topic so

that unstated assumptions can be recognized.

Module 1 AI Problems 8
 In addition to these mundane tasks, may people can

also perform one or may be more specialized tasks in

which carefully acquired expertise is necessary.

 Tasks including Engineering design, scientific discovery,

financial planning

 Programs that can solve problems in these domains

also fall under the protection of AI

Module 1 AI Problems 9
Mundane Tasks
 Perception
 Vision

 Speech

 Natural Language
 Understanding

 Generation

 Translation

 Commonsense Reasoning
 Robot Control

Module 1 AI Problems 10
Formal Tasks
 Games

 Chess

 Backgammon

 Checkers - Go

 Mathematics

 Geometry

 Logic

 Integral calculus

 Proving properties of programs

Module 1 AI Problems 11
Expert Tasks

 Engineering

 Design

 Fault finding

 Manufacturing, planning

 Scientific analysis

 Medical diagnosis

 Financial analysis
Module 1 AI Problems 12
 A person who knows how to perform tasks from several of
the categories learns the necessary skills in a standard
order.

 First, perceptual, linguistic, and commonsense skills are

learned.

 Later expert skills are acquired.

 It might seem to make sense that the earlier skills are easier
and thus more agreeable to computerized duplication
rather than are the later, more specialized ones.

Module 1 AI Problems 13
 For this reason, much of the initial AI work was

concentrated in these.

 But it turns out that this immature assumption is not right.

 Although expert skills require knowledge, that many of us

do not have, they often require much less knowledge than

do the more mundane skills and knowledge is usually

easier to represent and deal with inside programs.

Module 1 AI Problems 14
 As a result, the problem areas where AI is now flourishing

most as a practical discipline are primarily the domains

that require only specialized expertize without the

assistance of commonsense knowledge.

Module 1 AI Problems 15
 There are now thousands of programs called expert

systems in day-to-day operation through out all areas of

industry and govt.

 Each of these systems attempts to solve part or all of a

practical, significant problems that previously required

human expertize.

Module 1 AI Problems 16
 The important questions related to AI are

 What are the underlying assumptions about intelligence?

 What kinds of techniques will be useful to solve AI

problems?

 If what level of detail, if all, are trying to model human

intelligence?

 How will we know when we have succeeded in building a

intelligence?
Module 1 AI Problems 17
The Underlying Assumption

 AI lies in the physical symbol system hypothesis

 A physical symbol system consists of a set of entities, called

symbols, which are physical patterns that can occur as

components of another type entities called an expression

(symbol structure).

 Thus the symbol structure is composed of a number of

instances of symbols related in some physical way

Module 1 AI Problems 18
 At any instance of time, the system will contain as

collection of these symbol structures.

 Besides these structures, the system also contains a

collection of processes that operate on expressions to

produce other expressions: process of creation,

modification, reproduction and destruction.

Module 1 AI Problems 19
 A physical symbol system is a machine that produces

through time an evolving collection of symbol structures.

 Such system exists in a world of objects wider than just

these symbolic expressions themselves.

Module 1 AI Problems 20
 The physical symbol system hypothesis:- A physical symbol

system has the necessary and sufficient means for general

intelligent action.

 This hypothesis is only a hypothesis.

 There appears to be no way to prove or disprove it on logical

grounds.

 So it must be subjected to empirical validation

Module 1 AI Problems 21
 As it has become increasingly easy to build computing

machines, so it has become increasingly possible to

conduct empirical investigations of the physical symbol

system hypothesis.

 In each such investigation, a particular task that might be

regarded as requiring intelligence is selected.

 A program to perform the task is proposed and then tested.

Module 1 AI Problems 22
 The importance of the physical symbol system hypothesis

is twofold.

 It is a significant theory of the nature of human intelligence

and so is of great interest to psychologists.

 It also forms the basis of the believe that it is also possible

to build programs that can perform intelligent tasks now

performed by people.

Module 1 AI Problems 23
AI Technique

 The techniques that are appropriate for the solution of

AI problems.

 Intelligence requires knowledge.

 To compensate for its one overpowering asset,

indispensability, knowledge possesses some less

desirable properties

Module 1 AI Problems 24
 It is voluminous.

 It is hard to characterize accurately.

 It is constantly changing

 It differs from data by being organized in a way that

corresponds to the way it will be used.

 An AI technique is a method that exploits knowledge that

should be represented in such a way that:

Module 1 AI Problems 25
 The knowledge captures generalization:

 It is not necessary to represent separately each individual

situation.

 Instead, situations that share important properties are grouped

together.

 If knowledge does not have this property, excessive amounts of

memory and updating will be required.

 So we usually call something without this property data rather than

knowledge
Module 1 AI Problems 26
 It can be understood by people who must provide it:

 Although for many programs, the bulk of the data can

be acquired automatically, in many AI domains, most of

the knowledge a program has must ultimately be

provided by people in terms they understand.

Module 1 AI Problems 27
 It can easily be modified to correct errors and to reflect

changes in the world and in our world view.

 It can be used in a great many situations even if is not

totally accurate or complete.

 It can be used to help overcome its own sheer bulk by

helping to narrow the range of possibilities that must be

usually be considered

Module 1 AI Problems 28
 Although AI techniques must be designed in keeping with
these constraints imposed by AI problems, there is some
degree of independence between problems and problem
solving techniques.

 It is possible to solve AI problems without using AI

techniques.

 It is possible to apply AI techniques to solve non-AI

problems

Module 1 AI Problems 29
The Level of the Model

 Efforts to build programs that perform tasks the way people can

do can be divide into two classes

 Programs in the first class attempt to solve problems that do not

really fit our definition of an AI task.

 They are problems that a computer could easily solve, although

that easy solution would exploit mechanism that do not seem to

be available to people.

Module 1 AI Problems 30
 The second class of programs that attempt to model

human performance are those that do things that fall more

clearly within our definition of AI tasks; they do things that

are not trivial for the computer.

 There are several reasons one might want to model human

performance at these sorts of tasks:

Module 1 AI Problems 31
 To test psychological theories of human performance:

 To enable computers to understand human reasoning.

 To enable people to understand computer reasoning

 To exploit what knowledge we can glean from people

Module 1 AI Problems 32
Criteria for Success

 How will we know if we have succeeded?

 How will we know if we have constructed a machine

that is intelligent?

 In 1950, Alan Turing proposed aa method for

determining whether a machine can think.

 The method is known as Turing Test.

Module 1 AI Problems 33
 To conduct this test we need two people and the machine

to be evaluated.

 One person plays the role of the interrogator, who is in a

separate room from the computer and the other man.

Module 1 AI Problems 34
Module 1 AI Problems 35