Artificial Intelligence

Dr. Subhash Singh,

Associate Professor
Department of Mechanical and Automation Engg.
Indira Gandhi Delhi Technical University for Women
New Delhi, INDIA-110006
E-Mail:[email protected], [email protected]
Google Scholar Profile: https://scholar.google.com/citations?hl=en&user=qjQ8UTUAAAAJ
Web of Science Researcher ID: AAA-1824-2019
ORCID ID: 0000-0002-4874-4805
1
2
Syllabus:

3
4
 Goal of this course

 To introduce you to the field of Artificial Intelligence

 To explain the challenges inherent in building an “intelligent system”
 To explain the

 Key Paradigms
 Core Techniques
 Algorithms
On taking this course you should be able to:
 understand the role of basic knowledge representation
 Problem solving and learning methods in AI
 Access the applicability, strengths and weaknesses of these methods
in solving particular engineering problems.
 How to develop intelligent systems by assembling solutions to concrete
computational problem.
 Appreciate the role of problem solving, natural language processing and vision in
understanding human intelligence from a computational perspective.
Introduction to Artificial Intelligence
Instructional objective of the 1st UNIT:

 understand the definition of artificial intelligence

 Discuss the different components involved with intelligent behaviour
 Examine the different ways of approaching AI.
 Look at some example systems that use AI.
 Trace briefly the history of Ai.
What is artificial intelligence (AI)?

Artificial Intelligence

 is concerned with the design of intelligence in an artificial device.

 Term coined by McCarthy in 1956.

What is intelligence?

Humans?

 Behaves as intelligently as a human

Two schools of thoughts
 Behave in the best possible manner

 Thinking? Thinking intelligently

 Acting ? Acting intelligently

• What to look at:

Thought processes/reasoning Vs Behaviour

• How to measure performance:

Human like performance Vs Ideal performance
Approaches to AI:
Thought/reasoning

Systems that think like humans Systems that think rationally

(Turing Test) (Laws of thoughts/logic)

Human like
performance Ideal performance

Systems that think like humans Systems that think rationally

(Cognitive Science) (Rational Agents)

Behaviour
Methods/ Algorithms enabled by
constrains exposed by
representations that support models
targeted at thinking, perception and
action.
Typical AI problems:

• Intelligent entities (or agents) need to be able to do both “mundane task” and “expert task”.

Mundane Task:

 Planning route, activity

 Recognizing (through vision) people, objects
 Communicating (though natural language)
 Navigating round obstacle on the street

Expert Tasks:

 Medical diagnosis
 Mathematical Problem solving
 What are easy and what are hard task?

 It has been easier to mechanise many of the high level tasks we associate with “intelligence”
in people.
• Symbolic integration
• Proving theorems
• Playing chess
• Medical diagnosis

 It has been very hard to mechanise tasks that lots of animals can do
• Walking around without running into thing
• Catching prey and avoiding predators
• Interpreting complex sensory information
• Modelling the internal states of others animals from their behaviour

Intelligent Behaviour : Perception

Reasoning
Learning
Understanding language
Solving problem
1842
Turing Test:

17
The Turing test involves a computer, a human interrogator and a human foil. The
interrogator attempts to determine, by asking questions of the other two participants,
which is the computer. All communication is via keyboard and screen. The
interrogator may ask questions as penetrating and wide-ranging as he or she likes,
and the computer is permitted to do everything possible to force a wrong
identification. (So smart moves for the computer would be to say 'No' in response to
'Are you a computer?' and to follow a request to multiply one huge number by
another with a long pause and an incorrect answer.) The foil must help the
interrogator to make a correct identification. A number of different people play the
roles of interrogator and foil, and if sufficiently many interrogators are unable to
distinguish the computer from the human being then it is to be concluded that the
computer thinks.
 AI is one of the newest fields in science and engineering. Work started in
earnest soon after World War II, and the name itself was coined in 1956. Along with
molecular biology, AI is regularly cited as the “field I would most like to be in” by
scientists in other disciplines. A student in physics might reasonably feel that all the
good ideas have already been taken by Galileo, Newton, Einstein, and the rest. AI,
on the other hand, still has openings for several full-time Einsteins and Edisons.

19
AI currently encompasses a huge variety of subfields, ranging from the general
(learning and perception) to the specific, such as playing chess, proving
mathematical theorems, writing poetry, driving a car on a crowded street, and
diagnosing diseases. AI is relevant to any intellectual task; it is truly a universal field.

AI applications include:

 Advanced web search engines e.g., Google

 Recommendation systems used by YouTube, Amazon and Netflix

 Understanding human speech such as Siri and Alexa

 Self-driving cars e.g., Tesla

 Automated decision-making and competing at the highest level in

strategic game systems (such as chess and Go).
20
 Intelligent Agent

Objective: Define agent

Define a intelligent agent
Define a rational agent
Explain bounded rationality
Discuss different type of environment
Explain different agent architecture

21
Agent : Percepts

Ex. See (camera)

Sensors
Hear (Sonar sensor)
Percept sequence
Inputs-different inputs
from the environment

Environment

Agent

Actuators Change
Effectors

Actions

22
Agent :
 Agent operate in an environment
 Perceives its environment through sensors
 Act upon its environment through actuators/effectors
 Have goals
 Agent program implement mapping from percept sequence to actions
 Performance measure to evaluate agents
 Autonomous agent decide autonomously which action to take in the current
situation to maximize progress towards its goals

Percepts : is the inputs that an intelligent agent is perceiving at any given moment

23
Sensors and Effectors:

 A agent perceives its environment through sensors

 The complete set of inputs at a given time is called a percept

 The current percept or a sequence of percepts can influence the actions of an agent

 It can change the environment through effectors/actuators

 An operation involving an actuator is called an action

 Actions can be grouped into action sequences

24
 Performance
Behaviour and performance of Ais in terms of agent function

 Perception history (sequence) to action mapping

 Ideal mapping specifies which actions an agent should take at any point of moment/time

Performance measure: a subjective measure to characterize how successful an agent is

(ex. Speed, Power usage, accuracy, money etc.)

Examples of Agents:

Humans:
 Eyes, ears, skin, taste buds, etc for sensors
 Hands, fingers, legs, mouth for effectors/acuators
25
Robots:
 Camera, infrared, sonar sensor etc. for sensors
 Grippers, wheels, lights, speakers etc. for effectors/acuators

Xavier (CMU)

Xavier was designed and built by CMU (Carnegie Mellon University) graduate students in 1993
The overall goal was to develop autonomous mobile robots that can operate reliably over
extended periods of time, can learn and adapt to their environments, and can interact with
people in a socially acceptable manner.
26
 Cog Robot (MIT)

Cog was a project at the Humanoid Robotics Group of the Massachusetts Institute of Technology. It
was based on the hypothesis that human-level intelligence requires gaining experience from
interacting with humans, like human infants do.
27
Aibo Robot

AIBO is a series of robotic dogs designed and manufactured by Sony. Sony announced
a prototype Aibo in mid-1998, and the first consumer model was introduced on 11 May
1999. New models were released every year until 2006.
Software Agent: Softbot

 Functions as sensors
 Functions as effectors/acuators

Ex. Askjeeves.com, Indian railways (askdisha2.0), Microsoft team

etc.

Expert systems:

 Cardiologist Ex. Apple watch

29
Autonomous Spacecraft:

Ex. Mars rovers etc.

Intelligent Buildings:
Lightening conditions, air-conditioning etc.

30
 Fundamental faculties of intelligence

 Acting
 Sensing
 Understanding reasoning and learning

 In order to act you must sense. Blind action is not a characterization of intelligence

 In case of robots, sensing and acting, understanding is not necessary

 Sensing needs understanding to be useful

31
 Rational Agent
 AI is about building rational agent
 An agent is something that perceives and act
 A rational agent always does the right thing

o What are the functionality (goals)?

o What are the components?

o How do we build them?

Rationality
Perfect Rationality:

 Assumes that the rational agent knows all and will take the action that maximizes her utility.
 Human beings do not satisfied this definition of rationality.

32
Bounded Rationality: Herbert Simon 1972

 Because of the limitation of the human mind, humans must use approximate methods to handle
many tasks.

Rational Action:

 The action that maximizes the expected value of the performance measure given the percept
sequence to date.

 Rational= Best?

o Yes, to the best of its knowledge

 Rational = Optimal?

o Yes, to the best of its ability

o And its constraints

33
 Omniscience
 A rational agent is not omniscient

 It does not know the actual outcome of its actions

 It may not know the certain aspects of its environment

 Rationality must take into account the limitation of the agent

 Percept sequence, background knowledge, feasible actions

 Deal with the expected outcome of actions

34
 Bounded Rationality

 Evolution did not give rise to optimal agent but to agents which are some sences
locally optimal at best.
 In 1957, Herbert Simon proposed the notion of bounded rationality. The property of the agent
that behaves in the manner that is near optimal with respect to its goals as its resources
will allow.

35
 Agent Environment

 Environment in which an agent operate can be defined in different ways.

It is helpful to view the following definitions as reffering to the way the environment appears
From the point of view of the agent itself

Environment observability

 Full observable:
 All the environment relevant to the action being considered is observable
 Such environment are convenient, since the agent is freed from the task of keeping track of the
changes in the environment.

 Partially observable :
 The relevant features of the environment are only partially observable.

Examples: Fully observable-Chess, Partially observable-Poker

36
 Environment: Determinism
Deterministic : The next stage of the environment is completely described by
the current stage and the current action. Image analysis

Stochastic : If an element of interference or uncertainty occurs then the environment is

stochastic. Note that a deterministic yet partially observable environment will
appear to be stochastic to the agent. Ludo

Strategic : environment wholly determined by the preceding state and the action of multiple
agents is called strategic. Chess

Environment: Episodicity
Episodic / Sequence :

 An episodic environment means that subsequent episode do not depends on what actions occurred
In previous episodes

 In and sequential environment the agent engages in a series of connected episodes.

37
 Environment: Dynamism

Static Environment : does not change from one state to the next while the agent is considering
Its course of action. The only changes to the environment as those caused by the agent itself.

Dynamic Environment : changes over time independent of the actions of the agent-
and thus if an agent does not respond in a timely manner this counts as a choice to do nothing.

Environment: Continuity

Discrete/ Continuous :

If the number of distinct percepts and actions is limited, the environment is discrete otherwise
it is continuous

38
 Environment: Other agents

Single agent/ Multi agent :

 If the environment contains other intelligent agents, the agent need to be concerned
strategic, game-theoretic aspects of the environment (for either cooperative or
competitive agents)
 Most engineering environment do not have multi agent properties, whereas most social
economic systems get their complexity from the interactions of (more or less) rational agent

39
 Thank you
For Query please contact:
[email protected]
[email protected]
Phone No:9045704384

40