Transcript MACHINE INTELLIGENCE

Artificial Intelligence and Neural Networks
Humans:
Decision
Making
Process
Tools:
Computers and
IT. VB, VBA,
Excel, InterDev,
Etc.
DSS
Data:
Facts pertinent
to the decision
at hand.
Algorithms:
Math/Flow Chart
stuff that helps the
tools help the humans
make decisions.
MACHINE INTELLIGENCE
Will computers become as smart
as humans within the next 50
years?
IBM’S “DEEP BLUE” CHESS
PLAYING COMPUTER
A couple of years ago (1997), IBM’s
Deep Blue computer beat world chess
champion Gary Kasporov in a chess
match. Does that mean Deep Blue is
“smarter” than Kasporov when it
comes to playing chess?
IBM’S “DEEP BLUE” CHESS
PLAYING COMPUTER
What if I told you Deep Blue has to
look at a million times more scenarios
than Kasporov to settle on a move?
See http://www.ishipress.com/hamlet.htm
Raw power
Artificial Intelligence
• Artificial intelligence is behavior by a
machine that, if performed by a human
being, would be called intelligent
• "Artificial Intelligence is the study of how
to make computers do things at which, at
the moment, people are better" (Rich and
Knight [1991])
• AI is basically a theory of how the human
mind works (Mark Fox)
Objectives of
Artificial Intelligence
(Winston and Prendergast [1984])
• Make machines smarter (primary
goal)
• Understand what intelligence is
(Nobel Laureate purpose)
• Make machines more useful
(entrepreneurial purpose)
Signs of Intelligence
• Learn or understand from experience
• Make sense out of ambiguous or
contradictory messages
• Respond quickly and successfully to
new situations
• Use reasoning to solve problems
Signs of Intelligence
(cont’d)
Deal with perplexing situations
•
• Understand and Infer in ordinary,
rational ways
• Apply knowledge to manipulate the
environment
• Think and reason
• Recognize the relative importance of
different elements in a situation
Turing Test for Intelligence
A computer can be considered to be
smart only when a human
interviewer, “conversing” with both
an unseen human being and an
unseen computer, could not
determine which is which
AI Computing
• Based on symbolic representation and
manipulation
• A symbol is a letter, word, or number
represents objects, processes, and their
relationships
• Objects can be people, things, ideas,
concepts, events, or statements of fact
• Create a symbolic knowledge base
AI Computing (cont’d)
• Uses various processes to manipulate the
symbols to generate advice or a
recommendation
• AI reasons or infers with the knowledge
base by search and pattern matching
• Hunts for answers
(Algorithms often used in search)
Some interesting AI Web Destinations
AI software and FAQs
http://www.cs.cmu.edu/Groups/AI/html/faqs/ai/
(fairly techie)
American Association for Artificial Intelligence
http://www.aaai.org
(fairly general)
PC Artificial Intelligence magazine
http://www.pcai.com/pcai
(just right for OMIS 661, in my opinion)
The AI Laboratory at MIT:
http://www.ai.mit.edu
An Overview of
Neural Computing
• Constructing computers that mimic certain
processing capabilities of the human brain
• Knowledge representations based on
– Massive parallel processing
– Fast retrieval of large amounts of
information
– The ability to recognize patterns based
on historical cases
Neural Computing = Artificial Neural
Networks (ANNs)
Input
data
Dendrite
input wire
Neuron #1
Axon
(output wire)
Weight
W1,2
Neuron #2
Dendrite
Synapse
(control of flow of
electrochemical fluids
Data
signals
Neuron #3
FIGURE 17.3 Three Interconnected Artificial Neurons
Axon
Learning
Three Tasks (over-simplified)
1. Compute Outputs
2. Compare Outputs with Desired
Targets
3. Adjust Weights and Repeat the
Process
• Set the weights by either some rules
or randomly
• Set Delta = Error = actual output
minus desired output for a given set
of inputs
• Objective is to Minimize the Delta
(Error)
• Change the weights to reduce the
Delta
• Information processing: pattern
recognition
• Different learning algorithms
Benefits of
Neural Networks
• Usefulness for pattern recognition, learning,
classification, generalization and
abstraction, and the interpretation of
incomplete and noisy inputs
• Specifically - character, speech and visual
recognition
• Potential to provide some of human problem
solving characteristics
• Ability to tackle new kinds of problems
• Robustness
• Fast processing speed
• Flexibility and ease of maintenance
• Powerful hybrid systems
Limitations of
Neural Networks
• Do not do well at tasks that are not done well
by people
• Lack explanation capabilities
• Limitations and expense of hardware
technology restrict most applications to
software simulations
• Training times can be excessive and tedious
• Usually requires large amounts of training
and test data
Some interesting Neural Web
Destinations
Brainmaker
http://www.calsci.com
Neural Works Professional II Plus
Neuralware, Inc