How We

1
How Were Going toSolve the AI Problem

• Pedro Domingos
• Dept. Computer Science Eng.
• University of Washington

2
What is the AI Problem?

• Build robots that do every job humans do,as well
  as them or better.
• (Preferably much better.)

3
Why Havent We Solved It Yet?

• Because no one has really tried
• Everyone works on subproblems
• Because we dont have the hardware
• But this will soon change

4
A Phase Shift in AI Research

• In 10 years (give or take), hardware will reach
  the computational power of the human brain
• Then we can really start trying
• Progress is much faster when you work on the
  actual problem
• In the meantime Lay the groundwork

5
Ways to Solve AI
Approach Proponents Examples
Mother of all KBs Doug Lenat Cyc
Web mining Tom Mitchell Oren Etzioni WebKB KnowItAll
Retrace evolution Rod Brooks Genghis, Cog
Robot baby Paul Cohen Robot Baby
One Algorithm Geoff Hinton Jeff Hawkins Neural networks
6
Mother of All KBs

• Hypothesis We dont need no new discoveries
  just a lot of knowledge
• Empirical test Miserable failure
• Itll take tens of thousands of rules
• No, hundreds of thousands
• No, wait, more like millions
• Deduction is not enough!
• We need induction and uncertain reasoning
• Cycorp now realizes this
• And at least they tried

7
Web Mining

• Lets read the Web instead ofmanually inputting
  formal knowledge
• Pros
• Theres a lot of stuff in the Web
• Language is great window into intelligence
• Great application value in its own right
• Cons
• The Web sucks
• Language is built on top of vision,motor
  control, everyday life, etc.

8
Retracing Evolution

• Human intelligence is too hard.Build an insect
  first!
• Well, that turns out to be easy, and
  doesntbring us much closer to human
  intelligence
• Brooks got tired after Genghis pals,and went
  straight to Cog (which did nothing useful)
• Evolution is blindingly slow
• Subsumption architecture still seems likea good
  idea

9
Robot Baby

• Build a robot and let it learn like a baby
• Pros
• Guaranteed to work! (Existence proof)
• It solves the real problem
• Cons
• Is it overkill? (Intelligent ? Human)
• Do we really have to wait 10 years for it to grow
  up?
• Too much to try at once (start w. symbol
  grounding?)
• And we dont have the hardware

10
One Algorithm

• Hypothesis Neocortex is all one algorithm
• Pretty good empirical support so far
• It does everything learning, reasoning,
  vision,language, motor control, etc.
• Shortest path to AI Figure out what this
  algorithm is
• Reverse engineer the brain? Not necessarily
• Testbed Digit recognition? No!
• Algorithm has to work on many different problems
  without change

11
How About This?

• Build a Robot Baby
• Power it with One Algorithm
• Add stages one by one (Subsumption)
• Feed it Cyc
• And then have it Read the Web

12
It Takes a (Global) Village
Inference
CollectiveKnowledgeBase
Rules
Queries
Facts
Answers
Users
Contributors
Feedback
Outcomes
Learning
Richardson and Domingos, KCAP-2003
13
What Can We Do Now?

• Algorithms that work on any number of cores
• Solve two problems simultaneously
• Learning and reasoning
• Vision and robotics
• Language and common sense
• Then solve three
• Solve series of increasingly hard problems
• Dont get stuck in local optima
• If you have 80/20 solution, move on to next
  harder problem
• Stay off the bandwagons

14
Got the Hardware. Now What?

• Divide and conquer doesnt work for AI
• Gluing pieces together doesnt work(engineering
  hits complexity wall)
• We need the right language
• Mechanics Calculus
• Electromagnetism Differential operators
• Alternating current Complex numbers
• Digital circuit design Boolean logic
• AI Not there yet (but see Markov logic)

15
Three Simple Tests

• Youre not solving the AI problem if
• Your system doesnt work online
• Your system doesnt simultaneously process more
  than one type of information
• Your system doesnt process so much
  informationit needs a focus of attention
  mechanism
• Consciousness Lots of information well
  integrated online with a focus of attention

16
When Will We Solve AI?

• Common view Never
• Kurzweil, Moravec 25 years
• Both wrong
• Solving AI is a long-term project
• How do we make sure were making progress?
• How do we speed up progress?
• How do we keep up motivation (and funding)?