Artificial Intelligence

(CS F407)

Introduction

Chittaranjan Hota
Professor, Computer Science & Information Systems Department
BITS Pilani Hyderabad Campus, Hyderabad
[email protected]
08.01.2018
Why Artificial Intelligence?
• "AI will add 2.3 million jobs by 2020“. ~ Gartner
• Just as consumer appliances have microprocessors,
apps will have AI. ~ Gartner, Jan 2017.
• AI is the Next Industrial Revolution. ~ Wall Street Journal,
August 2017.
• Robots 'are here to give us a promotion,' not take away
jobs. ~Gartner, 2017.

• AI makes it easier for you to search for photos, breaking

down language barriers, or helping you get things done
with your own personal digital assistant. But it’s also
providing us with new ways of looking at old problems
and helping transform how we work and live.
(Source: ai.google/)
 Absentia-VR: Norah AI

Norah AI media engine consists of a vast neural network, which can be used by interactive
content designers to create games in a short time. The content creators can train the machine
with a movie and Norah AI will automatically generate a high-end game by learning from the
movie's script and the characters that inhabit it.
Niki.ai: Shop on chat using
NLP and ML
GAN: Two NNs contesting
with each other (2014)

Generative Adversarial networks(GANs):

transforming still images into brief animations
(GIFs) that simulate the future of the scene,
MIT’s AI lab
Google's artificial intelligence bot
thinks the purpose of life is 'to live
forever'
• Human: What is the purpose of life?
• Machine: To serve the greater good.
• Human: What is the purpose of living?
• Machine: To live forever.
• Human: Where are you now?
• Machine: I'm in the middle of
nowhere.
• Human: What is the purpose of dying?
• Machine: To have a life.
• Human: What is the purpose of
emotions?
• Machine: I don't know.
 AI, ML, Deep Learning
AI ML Deep learning
This work It is difficult to I do not know
(travel plan) is program this how to make
tedious and computer to this computer
requires lot of make it understand
problem understand what I know.
solving. Let me what I want it Let it create
do it through a to do. Let it its’ own
digital do it by itself structure and
machine using right figure out the
(Computer). structure and data for itself.
(1950, Turing a set of (2012
test) examples. onwards:
(1952, Arthur Google,
Samuel’s Facebook
checker etc.)
game)
Homo sapiens--man the
wise
What is Intelligence?
• Intelligence is the ability to think, to imagine, to create,
memorize, understand, recognize patterns, make
choices, adapt to change and learn from experience.

• Intelligence is a property of mind that encompasses

many related mental abilities, such as the capabilities
to:
• reason
• plan
• solve problems
• think abstractly
• understand ideas and language
• learn
 What is AI?
• AI is the study of how to make computers do things
which at the moment people do better.

Think like humans Think rationally

“cognitive science” => formalize inference process

E.g. General Problem (laws of thought)
Solver [Newell&Simon 61]

Act like humans Act rationally

E.g., Eliza, Turing test Also, bounded rationality.
ELIZA
(1964-66) MIT Artificial Intelligence Laboratory by Joseph Weizenbaum

• 1990s ALICE
• Loebner prize- win
$25,000 if you pass the
test (2017)

Turing (1950) "Computing machinery and

intelligence”

Predicted that by 2000, a machine

might have a 30% chance of fooling a
lay person for 5 minutes.

Achieved. (Siri!  )

https://chatbotsmagazine.com/how-to-win-a-turing-test-the-loebner-prize-3ac2752250f1

(Source: Wiki)
Course outline
• Search
• Blind and Heuristic search
• Game playing
• Planning
• Knowledge representation and reasoning
• Propositional logic
• Predicate logic
• Resolution by refutation
• Reasoning under uncertainty
• Natural language understanding
• Machine learning
• Supervised learning (= learning from examples)
• Reinforcement learning (= learning by reinforcement)
Evaluation Scheme
Sl Component & Nature Duration Weightage Date and
No. Time
1. Coding Assignments (Take Home, Three * 30% ---
Numbers)
2. Announced Quizzes (Two) 30 mins 10% ---
3. Mid sem Test (Closed Book) 1.5 hr 25% 06/03
3.30 – 5.00
4. Comprehensive Exam (Part Open) 3 hrs 35% 04/05 AN

Chamber Consultation hour: Tuesday (4-5 pm, B224)

Goals of AI

• Understanding of computational mechanisms

needed for intelligent behavior

• Intelligent connection of perception and action

• Solve knowledge-intensive tasks

• Enhance human-computer interaction

Applications of AI
• Game Playing
Total-war, Thief, Creatures,
Halo, Fear, Black and White,
Max Payne-3, Kill Zone-2
etc…
Pokemon Go…X

• Speech recognition

• Computer vision
 Continued…
• Financial Decision Making
Credit risk assessment by Indian banks,…
• Natural Language Understanding
how are you आप कैसे हैं

• Scheduling and Planning

Ford Automaker relieves burden of scheduling worker assignments

with AI software (2011)

• Robotics and Path planning

• Home appliances using fuzzy logic iRobot Packbot used

in Afghan and Iraq
wars

• And many more like healthcare, transportation, Bigdata etc…

Foundations of AI
• Mathematics: Logic, Probability theory,
temporal, …

• Neuroscience (Moore’s law states that

computers will have as many gates as humans
have neurons in 2020)

• Control Theory (sense/action loop)

• Linguistics (Children can create sentences they

have never heard before)
An AI Technique
 Board: 9-element vector
0 : blank, 1 : X , 2 : O
 Move table: 19,683 Rows of 9-element vectors
 Algorithm:

1. transform board vector from base 3 to 10

2. use (1) as the move table index
3. change the board by using the vector from (2)

Program 1
Comments:

 Advantages:
Efficient in terms of time, optimal game of tic-tac-toe in theory.
 Disadvantages:
space?
work?
error prone - move table
three dimension - 327, no longer work at all
Program 2

 Board: same as program1

but different digits for different symbols.
 Turn: game moves 1,2,3,.....
odd-numbered move : X
even-numbered move : O

Algorithm : 3 sub procedures

a. Make2: Board[5] or Board [2, 4, 6, or 8]
b. Posswin (p): 18 for p = X
50 for p = O
c. go (n) : Move to Board [n]
Strategy (Program 2)
 Turn=1 go (1)
 Turn=2 If board [5] is blank, go (5), else go (1)
 Turn=3 If board [9] is blank go (9), else go (3)
 Turn=4 If posswin (X) is not 0, go(Posswin(X)), else go(Make2)
 Turn=5 If posswin (X) is not 0, go(Posswin(X)), else if posswin (O)
is not 0, go (Posswin(O)), else if board [7] is blank, go(7), else
go(3)
 Turn =6 If posswin (O) is not 0, go (posswin (O)), else if posswin (X)
is not 0, go (Posswin (X), else go (Make2)
 Turn =7 If posswin (X) is not 0, go (posswin (X), else if posswin (O)
is not 0, go (posswin (O), else go (anywhere that is blank)
 Turn=8 If posswin (O) is not 0, go (Posswin (O)), else if (posswin
(X) is not 0, go (posswin (X)), else go (anywhere that is
blank)
Comments:

 Is it efficient than Program 1 (time)?

 Is it efficient than Program 1 (space)?
 More clarity (strategy)
 Easier to change (strategy)
 Cannot extend (generalize) to three dimension
Program 2'

 program 2 board
 magic square 15
 possible win check:
S = sum of two paired owned by a player
D = 15 – S
if 0 < D < 10 and Board [D] is empty then the player
can win
8 3 4
1 5 9
6 7 2
Program 3
Data structure: boardposition (a nine element vector representing the
board, a list of board positions that could result from the next move, and
a number representing how likely it is going to lead to a win)

Comments:
1. Require much more
time to consider all
possible moves.
2. Could be extended to
handle more
complicated games.

Img. Source: www.neverstopbuilding.com

Next: Problem solving using Heuristic search