HUMAN-LEVEL ARTIFICIAL INTELIGENCE & COGNITIVE SCIENCE Nils - - PowerPoint PPT Presentation

▶
human level artificial inteligence cognitive science
SMART_READER_LITE
LIVE PREVIEW

HUMAN-LEVEL ARTIFICIAL INTELIGENCE & COGNITIVE SCIENCE Nils - - PowerPoint PPT Presentation

Dec 05, 2022 298 likes •689 views

HUMAN-LEVEL ARTIFICIAL INTELIGENCE & COGNITIVE SCIENCE Nils J. Nilsson Stanford AI Lab http://ai.stanford.edu/~nilsson Symbolic Systems 100, April 15, 2008 1 OUTLINE Computation and Intelligence Approaches Toward HLAI The Current

slide-1
SLIDE 1

HUMAN-LEVEL ARTIFICIAL INTELIGENCE & COGNITIVE SCIENCE

Nils J. Nilsson Stanford AI Lab

http://ai.stanford.edu/~nilsson Symbolic Systems 100, April 15, 2008

1

slide-2
SLIDE 2

OUTLINE

Computation and Intelligence Approaches Toward HLAI The Current Situation

2

slide-3
SLIDE 3

THE COMPUTER: A UNIVERSAL MACHINE

“It can be shown that a single special machine of that type [a Turing machine] can be made to do the work of all. It could in fact be made to work as a model of any other machine. The special machine may be called the universal machine.” —Alan Turing “The importance of the universal machine is clear. We do not need to have an infinity of different machines doing different jobs. A single one will

  • suffice. The engineering problem of producing various machines for

various jobs is replaced by the office work of ‘programming’ the universal machine to do these jobs.” —Alan Turing “[Turing] decided the scope of the computable encompassed far more than could be captured by explicit instruction notes, and quite enough to include all that human brains did, however creative or original.” —Andrew Hodges, a Turing Biographer

3

slide-4
SLIDE 4

THE PHYSICAL SYMBOL SYSTEM HYPOTHESIS (PSSH)

“A physical symbol system [i.e., a computer] has the necessary and sufficient means for intelligent action.”

—Allen Newell and Herbert Simon

Relevance to Cognitive Science: Computational processes can explain (be a theory of) human intelligence. Relevance to AI: Computational processes can implement human intelligence.

But, it’s just a hypothesis!

4

slide-5
SLIDE 5

THE NAYSAYERS CLAIM (Among Other Things) “THE BRAIN IS NOT A COMPUTER!”

Computation: mainly serial 109 ops/sec 109 transistors digital/discrete (even binary!) disembodied silicon subject to crashes . . . The Brain: highly parallel 103 ops/sec 1014 neurons; 1017 synapses analog/continuous embodied protein fault-tolerant . . .

5

slide-6
SLIDE 6

WHAT IS A COMPUTER? WHAT IS A BRAIN?

Programs Registers, Machine Ops Logic Gates (AND’s, OR’s, ) 0’s and 1’s … Transistor Currents, Magnetizations Quantum Mechanics Cell Assemblies/Modules? Neurons, Axons, Dendrites Models of Neo-Cortex Perceptual/Motor Apparatus Human Intelligence Desires, Beliefs, Intentions? Mentalese? ??? Goals, Plans, Reactions? ???

CS EE Cog. Sci. Phys.

Plans, Goals, Inference, Logic Human-Level AI ? ? ?

AI

Depolarizations Neurotransmitters Genomic Activity Chemical Reactions

Neur. Sci. Bio. Chem.

Symbol Processing Data Structures (Lists, etc.) Graphical Models, “Blackboards”, Semantic Networks Neural Networks

6

slide-7
SLIDE 7

CAN COMPUTATIONAL SYSTEMS BE INTELLIGENT? HOW WOULD WE KNOW? The Turing Test The “Employment” Test

7

slide-8
SLIDE 8

Meeting and Convention Planner Maid and Housekeeping Cleaner Receptionist Financial Examiner Computer Programmer Roofer’s Helper Library Assistant Procurement and Sales Engineer Farm, Greenhouse, Nursery Worker Dishwasher Home Health Aide Small Engine Repairer Paralegal Lodging Manager Proofreader Tour Guide and Escort Geographer Engine and Other Machine Assembler Security Guard Retail Salesperson Marriage and Family Counselor Hand Packer and Packager

*From “America’s Job Bank,” a list of more than 1,500 jobs. Available at www.jobsearch.org/help/employer/SSONetJobCodeListbyCategory2.html

SOME JOBS THAT HUMANS PERFORM* CAN THEY BE AUTOMATED?

8

slide-9
SLIDE 9

9

slide-10
SLIDE 10

HOW TO PROCEED? LET’S LOOK AT SOME THINGS AI HAS TRIED

10

slide-11
SLIDE 11
  • Try to program some fundamental processes

thought to be involved in intelligence

  • Try to imitate the brain
  • Try to simulate the performance of ever more complex

biological organisms

  • Try to simulate biological evolution
  • Try to “educate” simple (child-like) programs to make

them more intelligent and capable

  • Try to program some activities thought to require

intelligence

11

THINGS AI HAS TRIED

slide-12
SLIDE 12
  • 1. Programming Activities That

Require Intelligence

Game playing Theorem proving Pattern recognition (images, speech, . . .) Natural language understanding Autonomous activity (robots that can perceive, plan and act) Expert judgment . . .

12

slide-13
SLIDE 13

Game-Playing

13

slide-14
SLIDE 14

Speech Recognition

14

slide-15
SLIDE 15

Autonomous Activity

15

slide-16
SLIDE 16

Autonomous Activity

16

slide-17
SLIDE 17

Expert Judgment

Part of a Large Bayesian Network Used for the Diagnosis of Hepatobiliary Diseases

17

slide-18
SLIDE 18

WOULD SOME COMBINATION OF ALL OF THESE SKILLS ADD UP TO HLAI?

18

slide-19
SLIDE 19
  • 2. Programming Processes Thought to

be Involved in Intelligence

Logical reasoning Probabilistic reasoning Search Image processing Knowledge representation Learning Syntactic analysis Planning . . .

19

slide-20
SLIDE 20

Logical Reasoning Resolution Theorem Proving Propositional Satisfiability (SAT)

20

slide-21
SLIDE 21

Probabilistic Reasoning Bayesian Belief Networks

“Why won’t the car start?”

21

slide-22
SLIDE 22

Search A*, Hill-Climbing Recursive Back-Tracking

22

slide-23
SLIDE 23

Image Processing

  • E. g., Edge Extraction

23

slide-24
SLIDE 24

Representing Knowledge Semantic Networks, Cyc, WordNet

24

slide-25
SLIDE 25

100

  • 100

90 90

Learning

  • E. g., Reinforcement Learning

25

80 80

slide-26
SLIDE 26

WOULD SOME COMBINATION OF PROCESSES LIKE THESE PRODUCE HLAI?

26

slide-27
SLIDE 27
  • 3. Trying to Imitate the Brain

Neural Networks Models of the Neo-Cortex

27

slide-28
SLIDE 28

Neural Networks

Sejnowski, T. J. and Rosenberg, C. R., Parallel networks that learn to pronounce English text, Complex Systems 1, 145-168 (1987).

Input: Text Versions of English Words Output: Sound Training: Change Weights to Make Sound More Correct

28

slide-29
SLIDE 29

Models of Neo-Cortex

Jeff Hawkins, Tom Dean, David Mumford, Geoff Hinton, . . .

Large Graphical Models

29

slide-30
SLIDE 30

Rod Brooks’s Creepy-Crawly Things

  • 4. Simulating Simple Biological Organisms

30

slide-31
SLIDE 31

Genetic Algorithms, Genetic Programming

  • 5. Simulating Biological Evolution

http://www.cs.northwestern.edu/~fjs750/netlogo/ final/gpdemo.html

Target-Seeking Demo

http://www.handshake.de/user/blickle/Truck/ index.html

Truck-Backing Demo

31

slide-32
SLIDE 32

Turing’s “Child Programme” Cassimatis’s “Cognitive Substrate” Lenat’s “CYC Bootstrapping”

  • 6. “Educating” Educable Programs

32

slide-33
SLIDE 33

WHICH OF THESE APPROACHES WILL BE SUCCESSFUL? WE’LL HAVE TO WAIT AND SEE!

33

slide-34
SLIDE 34

THESE EFFORTS HAVE PRODUCED A GROWING ARMAMENTARIUM OF TECHNICAL TOOLS IN THE MEANTIME:

34

slide-35
SLIDE 35

Here is a Partial List:

Bayesian Belief Networks Hidden Markov Models Kalman Filtering POMDP’s A* Global Search Hill-Climbing Local Search GA/GP Resolution Theorem Prvg. SAT Encodings/Solvers Semantic Networks Reinforcement Learning Neural Networks Backpropagation Support Vector Machines Blackboard Architectures Monte Carlo Methods Statistical Grammars Expectation Maximization Inductive Logic Programming Teleo-Reactive Programs Particle Filtering Model-Based Vision

Will They Help Us Achieve HLAI? Are More Tools Needed?

35

slide-36
SLIDE 36

THE CURRENT SITUATION

The Tools Are Being Used to Solve Problems in Several Fields: Biology Genomics Chemistry Medicine Aeronautics Geology Data Mining Business . . . But Little Work is Being Done Toward HLAI. Why?

36

slide-37
SLIDE 37

PETER HART’S QUESTION

(AI’s Progress)

Age Date ?? 1970 40

diagnostician

20 2005

hot-rod driver

10

10-year-old human

37