CMSC450 AI Tools
Introduction

AI history

Charts the eras of AI from 1943 to the present on one timeline with the AI winters and dated events from the lecture, alongside the foundations of AI, historical themes, and the state of the art.

  • Introduction to AI · slides 19–27

Foundations of artificial intelligence

Introduction to AI · slide 19
  1. 01Philosophy
  2. 02Mathematics
  3. 03Economics
  4. 04Neuroscience
  5. 05Psychology
  6. 06Computer engineering
  7. 07Control theory & cybernetics
  8. 08Linguistics

A brief history of artificial intelligence

Introduction to AI · slide 20
First AI winter: Late 1970sAI winterSecond AI winter: Late 1980s–early 1990sAI winterpresentInception, 1943–1956Inception1943–1956Early enthusiasm & expectations, 1952–1969Early enthusiasm & expectations1952–1969A dose of reality, 1966–1973A dose of reality1966–1973Expert systems, 1969–1986Expert systems1969–1986Return of neural networks, 1986–presentReturn of neural networks1986–presentProbabilistic reasoning and machine learning, 1987–presentProbabilistic reasoning and machine learning1987–presentBig data, 2001–presentBig data2001–presentDeep learning, 2011–presentDeep learning2011–present1950: Alan Turing proposes the Turing test11957: Herbert Simon: “within 10 years a computer would be chess champion”21958: The New York Times on the perceptron: “New Navy Device Learns by Doing”31973: The Lighthill Report evaluates the state of AI for the British Science Research Council41980: John Searle’s Chinese Room thought experiment52009: Discussion of AI in The New York Times takes off62013: Levesque, “On Our Best Behavior” (IJCAI 2013): Winograd schemas as a better Turing test72016: Winograd schema challenge at IJCAI: best system 58%, humans 90%8194019501960197019801990200020102020
  • Era (slide 20)
  • Continues to the present
  • AI winter (slide 23)
  • Date from the slides

Eras

Select an era, here or in the chart, to highlight it.

Dates from the slides

  1. 1950 Alan Turing proposes the Turing test Introduction to AI · slide 5
  2. 1957 Herbert Simon: “within 10 years a computer would be chess champion” Introduction to AI · slide 21
  3. 1958 The New York Times on the perceptron: “New Navy Device Learns by Doing” Introduction to AI · slide 22
  4. 1973 The Lighthill Report evaluates the state of AI for the British Science Research Council Introduction to AI · slide 23
  5. 1980 John Searle’s Chinese Room thought experiment Introduction to AI · slide 11
  6. 2009 Discussion of AI in The New York Times takes off Introduction to AI · slide 25
  7. 2013 Levesque, “On Our Best Behavior” (IJCAI 2013): Winograd schemas as a better Turing test Introduction to AI · slide 7
  8. 2016 Winograd schema challenge at IJCAI: best system 58%, humans 90% Introduction to AI · slide 10

A prediction

Introduction to AI · slide 21

“It is not my aim to surprise or shock you – but … there are now in the world machines that think, that learn and that create. Moreover, their ability to do these things is going to increase rapidly until – in a visible future – the range of problems they can handle will be coextensive with the range to which human mind has been applied. More precisely: within 10 years a computer would be chess champion, and an important new mathematical theorem would be proved by a computer.”

— Herbert Simon, 1957
Predicted within 10 years
Came true 40 years later

Prediction came true – but 40 years later instead of 10.

1958 The New York Times

Introduction to AI · slide 22

The New York Times · 1958

New Navy Device Learns by Doing

Psychologist Shows Embryo of Computer Designed to Read and Grow Wiser

… The embryo—the Weather Bureau’s $2,000,000 “704” computer—learned to differentiate between right and left after fifty attempts in the Navy’s demonstration for newsmen. …

Headline, subhead, and an excerpt of the clipping on the slide, a report on the perceptron.

Boom to bust: AI winters

Introduction to AI · slide 23

Late 1970s

First AI winter

  • Machine translation deemed a failure
  • Fall of connectionism (perceptron limitations)
  • Lighthill Report (published in 1973) was an evaluation of the current state of AI at that time written for the British Science Research Council

Late 1980s–early 1990s

Second AI winter

  • At the heart of the commercialization of AI were expert systems. These systems were handcrafted by surveying experts and creating “if-then” rule sets accordingly.

Historical themes

Introduction to AI · slide 24

  1. Boom and bust cycles

    Periods of (unjustified) optimism followed by periods of disillusionment and reduced funding

  2. Silver bulletism

    “The tendency to believe in a silver bullet for AI, coupled with the belief that previous beliefs about silver bullets were hopelessly naïve”

    (Levesque, 2013)

  3. Image problems

    AI effect: As soon as a machine gets good at performing some task, the task is no longer considered to require much intelligence

    AI as a threat?

Current AI boom

Introduction to AI · slides 25–26

E. Fast and E. Horowitz, Long-Term Trends in the Public Perception of AI, AAAI 2017.

The slides show two of its charts. The first, “Percentage of Articles in the NYT about AI”, covers 1986–2016 with pessimistic, optimistic, and total lines; discussion takes off around 2009. The second splits the articles by theme:

  • Singularity (positive)
  • Singularity (negative)
  • Decision making
  • Education
  • Work (positive)
  • Work (negative)
  • Healthcare
  • Military applications
  • Cyborg (positive)
  • Cyborg (negative)
  • Entertainment
  • Ethical concerns for AI
  • Loss of control
  • AI in fiction
  • Transportation
  • Lack of progress

State of the art in AI

Introduction to AI · slide 27
  • Autonomous vehicles
  • Legged locomotion
  • Autonomous planning and scheduling
  • Machine translation
  • Speech recognition
  • Recommendations
  • Game playing
  • Image understanding
  • Medicine
  • Climate science

Approaches to AI Places these applications in the four approaches on a shareable board.