CS 4100 Artificial al Intelligence or: Jan-Willem van de Meent In - - PowerPoint PPT Presentation

▶
cs 4100 artificial al intelligence
SMART_READER_LITE
LIVE PREVIEW

CS 4100 Artificial al Intelligence or: Jan-Willem van de Meent In - - PowerPoint PPT Presentation

Jan 02, 2024 161 likes •768 views

CS 4100 Artificial al Intelligence or: Jan-Willem van de Meent In Instructor ite: https://course.ccs.neu.edu/cs4100f19 Websit At Attribution : many of these slides are modified versions of those distributed with the UC Berkeley CS188 materials

slide-1
SLIDE 1

CS 4100 Artificial al Intelligence

In Instructor

  • r: Jan-Willem van de Meent

Websit ite: https://course.ccs.neu.edu/cs4100f19

At Attribution: many of these slides are modified versions of those distributed with the UC Berkeley CS188 materials Thanks to John DeNero, Dan Klein, and Peter Abbeel.

slide-2
SLIDE 2

Today ay’s plan an

  • What is AI? Plus a brief history
  • What this course is about and some logistics
slide-3
SLIDE 3
slide-4
SLIDE 4
slide-5
SLIDE 5
slide-6
SLIDE 6
slide-7
SLIDE 7
slide-8
SLIDE 8
slide-9
SLIDE 9
slide-10
SLIDE 10
slide-11
SLIDE 11
slide-12
SLIDE 12

Defining Artificial al Intelligence

  • Precise definitions are surprisingly elusive
  • Informally

ally: The discipline of creating intelligent algorithms

  • Here we’ve just offloaded the complexity into the term intelligent
  • In

In press: Any algorithm that makes predictions

  • AI often means the same thing as machine learning
  • Most machine learning is essentially computational statistics
slide-13
SLIDE 13
  • Humans are in some case very good at making

rational decisions, but certainly not perfect

  • Brains are very hard to reverse engineer!
  • “Brains are to intelligence as wings are to flight”
  • Lessons learned from the brain: memory and

simulation are key to decision making

The Brai ain as as Inspirat ation for AI

slide-14
SLIDE 14
slide-15
SLIDE 15
slide-16
SLIDE 16
slide-17
SLIDE 17
slide-18
SLIDE 18

Thi This Cour urse: e: Rat ational al approaches to thinking and acting

slide-19
SLIDE 19

Rat ational al decisions

Here we use the term rat ational al in a specific, technical manner

  • Rat

ational al: maximally achieving pre-defined goals

  • Rationality only concerns what de

decisions are made (not the thought processes underpinning them)

  • Goal

als are expressed in terms of the ut utility of ou

  • utc

tcom

  • mes
  • Being rat

ational al means max aximizing your expected utility

We may think of this view of AI as

Computational Rationality

slide-20
SLIDE 20

The imperative: maximize expected utility

slide-21
SLIDE 21

Par art I: Mak aking Decisions (Acting Rationally) Par art II: Reas asoning under Uncertai ainty (Thinking Rationally) Par art III: Ethics of AI Thr Throug ugho hout: ut: Applications and emerging research

To Topi pics cs in n Thi This Co Cour urse

slide-22
SLIDE 22

Age Agent-bas ased Approac aches to Mak aking Decisions

  • We will be concerned with designing ag

agents that

  • operate in some env

environm nment ent (very broadly interpreted)

  • maximize some notion of expected uti

utility ty

  • Typically there are pe

percept ptio ions available and ac actions we can take

  • The agent operates in stat

ates (e.g., locations)

  • The environment might be stat

atic (constant) or dynam amic (changing)

  • It also might be fu

fully lly or only par artial ally observable

slide-23
SLIDE 23

Vac acuum-clean aner world

The agent might be a Roomba

slide-24
SLIDE 24

Vac acuum-clean aner world

The agent might be a Roomba We have:

  • some perceptions via sensors
slide-25
SLIDE 25

Vac acuum-clean aner world

The agent might be a Roomba We have:

  • some perceptions via sensors
  • actions we can take (suck, move right, …)
slide-26
SLIDE 26

Vac acuum-clean aner world

The agent might be a Roomba We have:

  • some perceptions via sensors
  • actions we can take (suck, move right, …)
  • utility we’d like to maximize

(some combination of ”don’t break” and ”clean all floors”)

slide-27
SLIDE 27

Vac acuum-clean aner world

The agent might be a Roomba We have:

  • some perceptions via sensors
  • actions we can take (suck, move right, …)
  • utility we’d like to maximize

(some combination of ”don’t break” and ”clean all floors”)

  • states where are we on the floor?
slide-28
SLIDE 28

Rat ational al vac acuum clean aner?

slide-29
SLIDE 29

Rat ational al vac acuum clean aner

Here actions depend deterministically on states; i.e., this is a look up table. We will call such agents reflex ag agents.

slide-30
SLIDE 30

Rat ational al vac acuum clean aner

Here actions depend deterministically on states; i.e., this is a look up table. We will call such agents reflex ag agents. Ques Question: n: How would we evaluate such an agent?

slide-31
SLIDE 31

Defining a a Good Notion of Utility is Har ard

  • One meas

asure of performan ance: Amount of dirt cleaned in an eight-hour shift.

  • Pr

Problem: The agent can maximize this performance by cleaning the floor, then dumping out all the dirt, and then cleaning it again.

  • Better meas

asure of performan ance: Amount of time that floor is clean.

  • Ru

Rule of Thumb (from Textbook) k): It is better to design utility according to outcomes, than according to how we think the agent should behave.

slide-32
SLIDE 32

A A brie ief sele lectiv tive his istor tory of

  • f AI

AI

slide-33
SLIDE 33
  • 1940

1940-1950: 1950: Early y da days ys

  • 1943: McCulloch & Pitts: Boolean circuit model of brain
  • 1950: Turing's “Computing Machinery and Intelligence”

A brief an and selective history of AI

slide-34
SLIDE 34

Min Minsk sky

Built the first Neural Network computer in 1950

slide-35
SLIDE 35

Min Minsk sky

Built the first Neural Network computer in 1950 The same year that Turing proposed his test

slide-36
SLIDE 36

The Turi The Turing ng Test Test

Image credit: http://turing100.blogspot.com/2012/05/one-month-to-biggest-turing-test.html

slide-37
SLIDE 37
slide-38
SLIDE 38

https://xkcd.com/329/

slide-39
SLIDE 39
  • 1940

1940-1950: 1950: Early y da days ys

  • 1943: McCulloch & Pitts: Boolean circuit model of brain
  • 1950: Turing's “Computing Machinery and Intelligence”
  • 1950

1950—70: 70: Exc xcitement – Lo Look, Ma, a, no han ands!

  • 1950s: Early AI programs, including Samuel's checkers program,

Newell & Simon's Logic Theorist, Gelernter's Geometry Engine

  • 1956: Dartmouth meeting: “Artificial Intelligence” adopted
  • 1965: Robinson's complete algorithm for logical reasoning

A brief an and selective history of AI

slide-40
SLIDE 40

Initial al Successes: Checkers

slide-41
SLIDE 41

“machines will be capable, within twenty years,

  • f doing any work a man can do”
  • Herbert Simon, in 1965

Bold clai aims an and extreme optimism

slide-42
SLIDE 42

Initial al Successes: Toy Worlds an and Sear arch

slide-43
SLIDE 43
  • 1940

1940-1950: 1950: Early y da days ys

  • 1943: McCulloch & Pitts: Boolean circuit model of brain
  • 1950: Turing's “Computing Machinery and Intelligence”
  • 1950

1950—70: 70: Exc xcitement – Lo Look, Ma, a, no han ands!

  • 1950s: Early AI programs, including Samuel's checkers program,

Newell & Simon's Logic Theorist, Gelernter's Geometry Engine

  • 1956: Dartmouth meeting: “Artificial Intelligence” adopted
  • 1965: Robinson's complete algorithm for logical reasoning
  • 1970

1970—90: 90: Knowledg dge-ba based d appr pproaches

  • 1969—79: Early development of knowledge-based systems
  • 1980—88: Expert systems industry booms
  • 1988—93: Expert systems industry busts: “AI Winter”

A brief an and selective history of AI

slide-44
SLIDE 44
  • 1940

1940-1950: 1950: Early y da days ys

  • 1943: McCulloch & Pitts: Boolean circuit model of brain
  • 1950: Turing's “Computing Machinery and Intelligence”
  • 1950

1950—70: 70: Exc xcitement – Lo Look, Ma, a, no han ands!

  • 1950s: Early AI programs, including Samuel's checkers program,

Newell & Simon's Logic Theorist, Gelernter's Geometry Engine

  • 1956: Dartmouth meeting: “Artificial Intelligence” adopted
  • 1965: Robinson's complete algorithm for logical reasoning
  • 1970

1970—90: 90: Knowledg dge-ba based d appr pproaches

  • 1969—79: Early development of knowledge-based systems
  • 1980—88: Expert systems industry booms
  • 1988—93: Expert systems industry busts: “AI Winter”
  • 1990

1990—2012: 2012: Statistical appr pproaches

  • Resurgence of probability, focus on uncertainty
  • General increase in technical depth
  • Agents and learning systems… “AI Spring”?

A brief an and selective history of AI

slide-45
SLIDE 45
  • 1940

1940-1950: 1950: Early y da days ys

  • 1943: McCulloch & Pitts: Boolean circuit model of brain
  • 1950: Turing's “Computing Machinery and Intelligence”
  • 1950

1950—70: 70: Exc xcitement – Lo Look, Ma, a, no han ands!

  • 1950s: Early AI programs, including Samuel's checkers program,

Newell & Simon's Logic Theorist, Gelernter's Geometry Engine

  • 1956: Dartmouth meeting: “Artificial Intelligence” adopted
  • 1965: Robinson's complete algorithm for logical reasoning
  • 1970

1970—90: 90: Knowledg dge-ba based d appr pproaches

  • 1969—79: Early development of knowledge-based systems
  • 1980—88: Expert systems industry booms
  • 1988—93: Expert systems industry busts: “AI Winter”
  • 1990

1990—2012: 2012: Statistical appr pproaches

  • Resurgence of probability, focus on uncertainty
  • General increase in technical depth
  • Agents and learning systems… “AI Spring”?
  • 2012

2012—: E : Excitement – Lo Look, Ma, a, no han ands ag agai ain?

  • Big data, big compute, neural networks
  • Some re-unification of sub-fields
  • AI used in many industries

A brief an and selective history of AI

slide-46
SLIDE 46

AI / ML L is star arting to be everywhere

  • Advertising
  • Search engines
  • Route planning
  • Spam / fraud detection
  • Automated help desks
  • Product recommendations
slide-47
SLIDE 47

What at Can an AI Do?

Qu Quiz: Which of the following can be done at present?

  • Play a mean game of Jeopardy?
  • Drive safely along a curving mountain road?
  • Drive safely along Huntington Avenue?
  • Converse successfully with another person for an hour?
  • Play world champion level GO
  • Put away the dishes and fold the laundry?
  • Translate spoken Chinese into spoken English in real time?
  • Write an intentionally funny story?
slide-48
SLIDE 48

Co Computer Vision (Perception)

3-D Understanding

Facial Recognition Image Segmentation Pose Recognition

Source: TechCrunch [Caesar et al., ECCV 2017] [DensePose]

slide-49
SLIDE 49

https://cvdazzle.com

slide-50
SLIDE 50

Lan Languag age an and Speech

slide-51
SLIDE 51

AlphaG aGo beat ats World Cham ampion Le Lee Sedol

slide-52
SLIDE 52

Play aying Atar ari with Deep Q-lear arning

[Mnih et al., Nature 2015]

slide-53
SLIDE 53

Course Lo Logistics

slide-54
SLIDE 54

Par art I: Mak aking Decisions (Acting Rationally)

  • Fast search / planning
  • Constraint satisfaction
  • Adversarial and uncertain search

Par art II: Reas asoning under Uncertai ainty (Thinking Rationally)

  • Bayes’ nets
  • Decision theory
  • Machine learning / Deep Learning

Par art III: Ethics of AI Thr Throug ugho hout: ut: Applications and emerging research

  • Natural language, vision, robotics, games, …

To Topi pics cs in n Thi This Co Cour urse

slide-55
SLIDE 55

Co Course logistics

Homepag age: https://course.ccs.neu.edu/cs4100f19/ Bo Book: k: Artificial intelligence, a modern approach, 3rd edition We may also draw from Sutton & Barto’s free book on reinforcement learning. Grad ading: 10% in-class exercises 45% homeworks and projects 30% midterm and final 15% final project

slide-56
SLIDE 56

In Instruction vs Assessment

Ins Instruct uction

  • In-class exercises
  • Homework
  • Projects

As Assessme ment

  • Midterm Exam
  • Final Exam
  • Final Project
slide-57
SLIDE 57

Ho Homewo eworks rks an and Projects

Ho Homew eworks

  • Due every week (approximately)
  • Electronic via grad

adescope

(we may also have a written components)

Pr Projects

  • Due every two weeks (approximately)
  • Based on Berkeley CS 188 materials
  • Submission in Python via grad

adescope

  • We will provide an autograder

(but will perform additional tests when grading)

  • Do not copy online solutions
slide-58
SLIDE 58

Ques Questions ns fo for me? e?

slide-59
SLIDE 59

To ToDo’s (by Monday ay)

  • Check out the website: https://course.ccs.neu.edu/cs4100f19/
  • Make sure you are registered on piazza and gradescope.
  • Complete HW0 (math diagnostic) and P0 (python tutorial)
  • Does not count towards final grade

(but will help you make sure you are prepared)

  • Do the readings for the next lecture