COMP 424 - Artificial Intelligence

What is Artificial Intelligence?

Instructors:
Jackie CK Cheung ([email protected])
Bogdan Mazoure ([email protected])
 Outline for Today
• Biological and artificial intelligence
• Overview of AI history
• Examples of AI applications

2
What is Intelligence?

3
 Possible Aspects of Intelligence
• Acquire, retain, and apply knowledge
• Apply logic and reason
• Be able to change and manipulate one’s environment
• Be able to adapt, foresee, and plan
• Be able to deal with uncertainty
• Possible modalities of intelligence:
• Spatial, verbal, logical, musical, social, …

4
 One Model: Biological Intelligence
• Sensory processing:
• Visual cortex.
• Auditory cortex.
• Somatosensory cortex.
• Motor cortex.
• Cognitive functions:
• Memory.
• Reasoning.
• Executive control.
• Learning.
• Language.

5
 Biological Intelligence
• A mix of general-purpose and special-purpose algorithms.
• General-purpose:
• Memory formation, updating, retrieval.
• Learning new tasks.
• Special-purpose:
• Recognizing visual patterns.
• Recognizing sounds.
• Learning language.
• All are integrated seamlessly!

6
What is AI?

7
 What is AI?
Human intelligence:
• Sensory processing:
• Visual cortex
• Auditory cortex
• Somatosensory cortex
• Motor cortex
• Cognitive functions
• Memory
• Reasoning
• Executive control
• Learning
• Language

8
 What is AI?
Human intelligence: Artificial Intelligence:
• Sensory processing:
• Visual cortex → Computer vision
• Auditory cortex → Signal/speech processing
• Somatosensory cortex → Haptics
• Motor cortex → Robotics
• Cognitive functions
• Memory → Knowledge representation
• Reasoning → Search, inference
• Executive control → Planning, decision-making
• Learning → Model learning
• Language → Language understanding

9
 What is AI?
Human intelligence: Artificial Intelligence:
• Sensory processing:
• Visual cortex → Computer vision
• Auditory cortex → Signal/speech processing
• Somatosensory cortex → Haptics
• Motor cortex → Robotics
• Cognitive functions
• Memory → Knowledge representation
• Reasoning → Search, inference
• Executive control → Planning, decision-making
• Learning → Model learning
• Language → Language understanding

10
 Possible Goals of AI
• Modeling or replicating human cognition using computers.
• Replicating human behaviours using computers
• Studying problems that others don’t know how to solve.
• Cool stuff!
• Game playing, machine learning, data mining, speech recognition,
computer vision, web agents, robots
• Useful stuff!
• Medical diagnosis, fraud detection, genome analysis, object
identification, space shuttle scheduling, information retrieval

11
 Goals of AI

Thinking Thinking
Humanly Rationally

Acting Acting
Humanly Rationally

12
 Different Goals of AI
What is one major obstacle to investigating these goals?

Thinking Thinking
Humanly Rationally

Acting Acting
Humanly Rationally

13
 Different Goals of AI

Thinking Thinking
Humanly Rationally
This would be
pretty awesome!
Acting Acting
Humanly Rationally

14
 Acting Humanly
• AI is about duplicating what the (human) brain DOES.
• Alan Turing (1912-1954) had interesting thoughts about this.
Can a machine think? -> If it could, how would we tell?
Turing (1950): “Computing machinery and intelligence”

Human interrogator ? Human

AI agent

An operator interacts with either the human or the AI agent.

Can he correctly guess which one?

15
 Turing’s Prediction
• By 2000, a machine would have a 30% chance of fooling a
lay person for 5 minutes.
• This actually happened in 2014:
http://www.bbc.com/news/technology-27762088
• Does this mean that we have solved AI?

• Suggested major components of AI:

• Knowledge representation, automated reasoning, language
understanding, machine learning

16
 AI = Acting Humanly?
• Humans have biological resource constraints
• Limited memory, thinking speed, attention span…
• Humans are often irrational
• Or else we do not understand our own notion of rationality
• e.g., don't act towards our own goals; hold contradictory beliefs
or preferences
• Kahneman and Tversky demonstrate ways that people are
systematically irrational.

17
 Optical Illusions
• Do we really want AI systems to replicate all of the
features of human perception and cognition?

• E.g., visual system:

https://michaelbach.de/ot/mot-breathingSquare/index.html

https://en.wikipedia.org/wiki/Ponzo_illusion

18
Different Goals of AI

Thinking Thinking
Humanly Rationally
Let’s try this!

Acting Acting
Humanly Rationally

19
 Acting Rationally
• Rational behaviour = doing the “right” thing.

• Doing what is expected to maximize goal achievement,

given the available information and available resources.

This is the flavour of AI we will focus on.

20
 Finally, Our Working Definition of AI
• Developing models and algorithms that can produce
rational behaviours in response to incoming stimuli and
information.

21
 Rational Agents
• This course is about designing rational agents.
• An agent is an entity that perceives and acts.
• Goal: Learn a function mapping percept histories to actions:
f : Ph → A
• A rational agent implements this function such as to maximize
performance.
• Performance measures: goal achievement, resource consumption, ...
• Caveat: Resource constraints (time, space, energy,
bandwidth, …) which make perfect rationality unachievable
• Objective: Find best function for given information and
resources

22
 Course Topics of COMP 424
• Search Basic tools
• Game playing
• Logical reasoning Logical
• Classical planning representations
• Probabilistic reasoning
• Learning probabilistic models
• Causal probabilistic models Probabilistic
• Reasoning with utilities representations
• Sequential reasoning and decision-making. Utility theory
• Applications

23
 AI Beginnings
• ENIAC: First super-computer, created in 1946.
• Early work in 1950s:
• Rosenblatt’s perceptron
• Samuel’s checkers player

24
Dartmouth Conference (1956)

25
 Dartmouth Conference (1956)
• Some of the attendees:
• John McCarthy: LISP, time-sharing, application of logic
to reasoning
• Marvin Minsky: popularized neural networks and
showed their limits, introduced slots and frames
• Claude Shannon: information theory, juggling machine
• Allen Newell and Herb Simon: bounded rationality,
general problem solver, SOAR

• The meeting coined the

term “artificial intelligence''

26
 Early AI Hopes and Dreams
• Make programs that exhibit similar signs of intelligence as
people: prove theorems, play chess, have a conversation.

• Logical reasoning was key.

• Learning from experience was considered important.
• The research agenda was geared towards building general
problem solvers.

• There was a lot of hope that natural language could be

easily understood and processed.

27
 AI Downswings
• Early successes did not scale up!
• Demos were impressive, but only worked in a narrow domain
• e.g., Machine translation for general texts

1966 Perceived failure of machine translation

1973, 1974 Major cut in AI research funding
1987-1993 “AI Winter”: many companies working in AI
fail

• Much progress actually made in this period, but

overpromising of results led to disappointments

28
 Recent AI: Statistics to the Rescue!
• Heavy use of probability theory, decision theory, statistics.
• Trying to solve specific problems rather than aim for general
reasoning.
• AI today is a collection of sub-fields:
• Perception and computer vision.
• Natural language understanding.
• Robotics
• Etc.
• Reasoning is now the part named “AI”.
• A lot of progress was made in this way!
• Some recent efforts try to put all this together

29
 AI system (1997): Chess playing
IBM Deep Blue defeats Garry Kasparov.
• Perception: advanced features of the
board.
• Actions: choose a move.
• Reasoning: search and evaluation of
possible board positions. www.bobby-fischer.net

http://www-03.ibm.com

30
 AI system (2008): Poker playing
University of Alberta’s Polaris defeats some
of the world’s best online pros.
• One variety of poker: Heads-up limit Texas
Hold’em (two players, limited betting
amounts)
• Perception: features of the game.
• Actions: choose a move.
• Reasoning: search and evaluation of
possible moves, machine
learning.
http://poker.cs.ualberta.ca/

31
AI system (2011): Jeopardy!

32
 AI system (2015): Atari

Similar video on YouTube:

https://www.youtube.com/watch?v=V1eYniJ0Rnk&ab_channel=TwoMinutePapers
33
AI system (2017): Self-Play for Go

34
 AI system (2021): AlphaFold
Predict protein structure from amino acid sequences
(Partial) solution to a 50 year old problem

Source: Nature
35
So, is AI solved?

36
37
 Online Demo Released

• https://app.inferkit.com/demo

38
 AI system (2014): City driving
Google cars have logged over 700,000 miles in autonomous
mode
• Sensors and actuators similar to Stanley (GPS, 3D laser point cloud,
cameras, odometry)
• Significant prior knowledge: City of Mountain View (12-square-mile)
is fully mapped at high resolution.

39
40
 How Do We Trust AI?
• Need to temper the hype!
• What kinds of problems do current AI technology solve?
• What are their limitations?

• How can we trust AI, when they can fail dramatically and
unexpectedly?

41