Soft Computing, Machine Intelligence

1
Soft Computing, Machine Intelligence Human
Intelligence

• Lluís A. Belanche
• belanche_at_lsi.upc.edu

2
Overview

• Soft Computing its components
• On intelligence
• Words are important!
• The problem of induction
• The elements of learning
• (whatever this is)
• Hybrid methods

3
What could SOFT COMPUTING be?

• To explicitly represent and exploit
• imprecision, uncertainty, robustness, data
  dependencies, learning and/or optimization
  ability, to achieve a working solution to a
  problem which is hard to solve.
• To find an exact (approximate) solution to an
  imprecisely (precisely) formulated problem.

4

• Parking a Car (difficult or easy?)

• Generally, a car can be parked rather easily
  because the final position of the car is not
  specified exactly. It it were specified to
  within, say, a fraction of a millimeter and a few
  seconds of arc, it would take hours of
  maneuvering and precise measurements of distance
  and angular position to solve the problem.
• ? High precision carries a high cost.

5
So what is the aim?

• The challenge is to put these capabilities into
  use by devising methods of computation which lead
  to an acceptable solution at the lowest possible
  cost.
• This should be the guiding principle of soft
  computing.

6
On intelligence

• What is Intelligence?
• What is the function of Intelligence?
• ? to ensure survival in nature
• What are the ingredients of intelligence?
• Perceive in a changing world
• Reason under partial truth
• Plan prioritize under uncertainty
• Coordinate different simultaneous tasks
• Learn under noisy experiences

7
Fundamental questions for intelligence ... or not
so ?

• Whats a computer?
• a simulation machine
• to simulate ? to replicate ? to be
• The frame problem
• The symbol grounding problem
• The Learning problem
• The Turing test (a problem for AI?)

8
Relation to AI

• Is soft computing rather than hard computing the
  foundation for Artificial Intelligence?
• Strong AI versus Weak AI a debate (Searle 80)
• Weak AI the appropiately programmed computer is
  just a program.
• Strong AI the appropiately programmed computer
  is more than a program, is a mind.

9
The primordial soup
Fuzzy Logic

• Belief
• Networks

Neural Networks
Soft Computing
Chaos Fractals
Evolutionary Algorithms
Rough Sets
10
Different methods different roles
Fuzzy Logic the algorithms for dealing with
imprecision and uncertainty Neural Networks
the machinery for learning and function
approximation with noise Evolutionary Algorithms
the algorithms for reinforced search and
optimization
uncertainty arising from the granularity in the
domain of discourse
11
Examples of soft computing

• TSP 105 cities,
• accuracy within 0.75, 7 months
• accuracy within 1, 2 days
• Compare
• absoulute best for sure with very good with
  very high probability

12
Are you one of the top guns?

• Consider
• Search space of size s
• Draw N random samples
• What is the probability p that at least one of
  them is in the top t ?
• Answer p 1 (1-t/s)N
• Example s 1012, N100.000, t1.000
• ? 1 in 10.000 !

13
On Algorithms

• Is it worth?

Specialized algorithms best performance for
special problems Generic algorithms good
performance over a wide range of problems
Generic Algorithms
Efficiency
Specialized Algo.
P
Problems
14
Words are important !

• What is a theory ?
• What is an algorithm ?
• What is an implementation ?
• What is a model ?
• What does non-linear mean ?
• What does non-parametric mean ?

15

• Learning Foreignia (Poggio Girosi93)

Can a machine learn to pronounce? 1. Do
nothing and wait 2. Learn all the pronunciation
rules 3. Memorize the pronunciation pairs 4.
Pick a subset of pronunciation pairs
and learn/memorize them 5. Pick a subset of
pronunciation pairs and develop a model
that explains them
16
The problem of induction

• Classical problem in Philosophy
• Example 1,2,3,4,5,?
• A more through example JT

17
What are the conditions for successful learning?

• Training data (sufficiently) representative
• Principle of similarity
• Target function within capacity of the learner
• Non dull learning algorithm
• Enough computational resources
• A correct (or close to) learning bias

18
Why NOT Integrations ?
Fuzzy Logic ANN ANN GA Fuzzy Logic ANN
GA Fuzzy Logic ANN GA Rough Set
Neuro-fuzzy hybridization is the most
visible integration realized so far.
19
Example 1 Fuzzy EAs

• Fuzzy rulebase for the dynamic control of an
  evolutionary algorithm

If D(Xt) is LOW then pmut is HIGH If f (Xt) is
LOW and D(Xt) is HIGH then Recomb is R1 . . .
20
Rough Sets
Z. Pawlak 1982, Int. J. Comp. Inf. Sci.

• Fundamental principle is to discover
  redundancies and dependencies between the given
  features of data to be classified.
• Approximate a given concept both from below and
  from above, using lower and upper approximations.
  
• Rough set learning algorithms can be used to
  obtain rules in IF-THEN form.

21
Rough Sets
Upper Approximation BX
Set X
Lower Approximation
xB (Granules)
.
x
xB set of all points belonging to the same
granule as of the point x
in feature space WB.
xB is the set of all points which are
indiscernible to point x in terms of feature
subset B.
22
Approximations of the set
w.r.t feature subset B
B-lower BX
Granules definitely belonging to X
B-upper BX
Granules definitely and possibly belonging to X
If BX BX, X is B-exact or B-definable Otherwise
it is Roughly definable
23
Example 2 Rough Fuzzy

• Fuzzy Set theory assigns to each object a degree
  
• of membership to represent an imprecise or
• vague concept.
• Rough Set theory focuses on the ambiguity
• caused by limited discernibility of objects.

Rough sets and Fuzzy sets can be integrated to
develop a stronger model of uncertainty.
24
Rough Fuzzy Hybridization A New Trend in
Decision Making, S. K. Pal and A. Skowron (eds),
Springer-Verlag, Singapore, 1999
25
And the Oscar goes to

• The real problem is not whether machines think,
  but whether men do.
• B.F. Skinner,
• Contingencies of Reinforcement