A Gentle Introduction to Computer Science for Artificial Intelligence

by

David Nader P.

2019

Main Concepts in CS: Algorithms, Machines, and Systems

Milestones in Computer Science

Active (inter-disciplinary) fields

Artificial Intelligence

Computer Science Influencers

Main Concepts in CS: Machines, Systems, and Algorithms

Main Concepts in CS

Etymology: Computer -> Computation -> Counting

Ref: https://en.wikipedia.org/wiki/Abacus#/media/File:Boulier1.JPG

Abacus: 2700–2300 BCE in Sumer

Quipu: First Millenium CE in the Andes

Main Concepts in CS

Computer Science is:

theory
engineering
experimentation

of Machines!

Main Concepts in CS

1) Machine: Enabling computations

Ref: https://en.wikipedia.org/wiki/Turing_machine_gallery#/media/File:Turing_machine_1.JPG

Main Concepts in CS

2) System: components and subcomponents interacting each other for a goal

Main Concepts in CS

Computer Science is:

theory
engineering
experimentation

of Machines!

Main Concepts in CS

3) Algorithm: Muhammad ibn Musa al-Khwarizmi

What should I do to withdraw money from an ATM?

Look for an ATM
Identify the card input
...
...
1. 1 Select the amount...

Main Concepts in CS

Computer Science Influencers

Ada Lovelace and Charles Babbage

Analytical Engine - 1810s

Bernulli Numbers Program - 1840s

Can Machines Think?

A.M Turing - 1950

Computer Science Influencers

Alan Turing

Cryptoanalysis
Computational Biology
Turing Test
Artificial Machines

Computer Science Influencers

Konrad Zuse and Akira Nakashima (switching circuit theory)

Zuse-z3: First Fully Automatic Digital Computer (1949)

Computer Science Influencers

John von Neumann and the von Neumann architecture

Game Theory
Set theory
Lattice Theory
Math formulation of Quantum Mechanics
Entropy
Mathematical Economics
Fluid Dynamics
Cellular Automaton, DNA and Universal Constructor

Computer Science Influencers

"Programming a Computer for Playing Chess" -Shannon

50

"Computing Machinary and Intelligence" -Turing

55

"Artificial Intelligence" workshop -McCarthy

"Logic Theorist" first AI program -Newell, Simon, & Shaw

59

"Machine Learning"-Samuel

84

97

09

11

17

18

Computer Science Influencers

"Programming a Computer for Playing Chess" -Shannon

50

"Computing Machinary and Intelligence" -Turing

55

"Artificial Intelligence" workshop -McCarthy

"Logic Theorist" first AI program -Newell, Simon, & Shaw

59

"Machine Learning"-Samuel

65

-Weizenbaum

84

"AI winter"-Shanck & Minsky

DeepBlue vs Kasparov

97

09

11

17

18

Computer Science Influencers

"Programming a Computer for Playing Chess" -Shannon

50

"Computing Machinary and Intelligence" -Turing

55

"Artificial Intelligence" workshop -McCarthy

"Logic Theorist" first AI program -Newell, Simon, & Shaw

59

"Machine Learning"-Samuel

65

-Weizenbaum

84

"AI winter"-Shanck & Minsky

DeepBlue vs Kasparov

97

09

Driverless Car by Google

11

Virtual Assistant -Siri by Apple

17

AlphaGo by

DeepMind

Dialog Agents (learning how to negotiate)

18

Turing Award to Deep Learning work

Computer Science Influencers

ACM Turing Award 2018: For conceptual and engineering breakthroughs that have made deep neural networks a critical component of computing

Yoshua Bengio

Geoffrey Hinton

Yann LeCun

Milestones in Computer Science

Analytical engine

1837: Mechanical general-purpose computer

Ref: https://en.wikipedia.org/wiki/Analytical_Engine#/media/File:AnalyticalMachine_Babbage_London.jpg

Milestones in Computer Science

Binary Logic

1854: Boolean Algebra

Ref: https://en.wikipedia.org/wiki/George_Boole

Milestones in Computer Science

Enigma

Crypto machine of German military during WW2

Ref: https://en.wikipedia.org/wiki/File:Enigma.jpg

Milestones in Computer Science

Digital Revolution

50. - 70s: Third Industrial Revolution
Internet users: 3.9 billion (49.5% of world population in 2016)

Ref: https://en.wikipedia.org/wiki/File:Internet_map_1024.jpg

Milestones in Computer Science

Programming Languages

Punched cards (40s)
Machine Language (1000111)
High-Level Language (e.g., java, c, c++, cobol, scala)

Milestones in Computer Science

Computer Graphics

60s: Verne Hudson and William Fetter of Boeing
Computer-generated imaginary (CGI)

Ref: https://en.wikipedia.org/wiki/Computer_graphics#/media/File:Blender_2.45_screenshot.jpg

Milestones in Computer Science

Artificial Intelligence

50s: Newell, Simon, McCarthy, Minsky, Samual
Intelligence demonstrated by Machines

Milestones in Computer Science

Object Recognition

CNNs-2012

Artificial Intelligence

Machine Learning

Deep Learning

Analytical Programming

Having the inputs, outputs, and the conditions to compute them

Artificial Intelligence

Having the inputs and outputs, but the conditions are unknown

Analytical Programming

Cats

Dogs

f(x)

George

Processing Features

Identify light/dark pixels

Identify edges

Identify combinations of edges

REF: https://www.pnas.org/content/pnas/116/4/1074.full.pdf

Identify facial

features

Recognition

Coding Features

Cats

Dogs

f(x)

George

Image Processing

f(x)

Cats

Dogs

George

Problem:

Manual

Feature Extraction

Solution:

Represenational Learning

Universal Approximation Theorem

f(x)

\approx

Neural Networks + Gradient Descent + GPU

p(x|y,\theta)

Deep Learning (classifier)

Cats

Dogs

f(x)

Deep Learning (training)

Cats

Dogs

f(x)

Dogs

Backpropagation

Deep Learning (inference)

Cats: 5%

f(x)

Dogs: 95%

"DL is essentially a new style of programming -differentiable programming- and the field is trying to work out the reusable constructs in this style. We have some: convolution, pooling, LSTM, GAN, VAE, etc" -- Thomas G. Dietterich

DL Applications

Image Generation

Face2Face or Facial Reenactment

Video Generation

Facial Recognition

Object Detection

Sentiment Analysis

Self-Play

Image Classification

Driverless Cars (Planes, Boats, Drones....)

Natural Language Processing

Neural Machine Translation

Limits of AI (approach)

Adversarial Attacks

"The Book of Why" -Judea Pearl (2018)

Main AI Techniques

"The Book of Why" -Judea Pearl (2018)

Judea Pearl (2018)

Reinforcement Learning

Judea Pearl (2018)

Machines are not able to compute retrospections (yet!)

"General Artificial Intelligence"

Neural Nets are Complex Systems

Interpretability: Why is NN producing those results?
Understandability: Why does deep learning work so well?
Towards Biologically Plausible Deep Learning

Limits of AI (ethics)

Job Automation: Amazon Go

Self-awareness

Autonomous

Weapons

AI- Bias

No privacy and Total Surveillance

Active (inter-disciplinary) fields

Medicine

Computer vision identifies signs of early Alzheimer's up to 6 years before clinical diagnosis

Active (inter-disciplinary) fields

Chemistry

"(...) predicting the material properties for new material candidates ... undiscovered compounds (...)"

Active (inter-disciplinary) fields

Physics

Finding patterns on LHC data

Determining particle properties (Higgs boson)

Active (inter-disciplinary) fields

History

"(...) humanity researchers can use these AI-parsed texts to search for specific words in a book, see how a popular narrative changed over time, analyze the evolution of the language ... (...)"

Active (inter-disciplinary) fields

Arts

Active (inter-disciplinary) fields

Arts

Edmond de Belamy, from La Famille de Belamy
generative Adversarial Network print, on canvas, 2018, signed with GAN model loss function in ink by the publisher, from a series of eleven unique images, published by Obvious Art, Paris, with original gilded wood frame
S. 27 ½ x 27 ½ in (700 x 700 mm.)