Transcript AI & Expert Systems

What is Artificial Intelligence?
• AI is the effort to develop systems that can
behave/act like humans.
• Turing Test
• The problem = unrestricted domains
–
–
–
–
human intelligence vastly complex and broad
associations, metaphors, and analogies
common sense
conceptual frameworks
Elements of AI
•
•
•
•
Natural Language Processing
Robotics
Perceptive Systems (Vision)
Expert Systems
How are Machines Intelligent?
• Constrained Heuristic Search
– How do you play chess?
• first move = 20 possible
• second move = 400 possible
• 7th move = 1,280,000,000 possible
– Depth First vs. Breath First Searching
– Ability to Learn
Decision Tree
Depth First Search
Breath First Search
Expert Systems
• Capture knowledge of an expert.
• Represent Knowledge as a
– rule base
• if then rules
– semantic net
• hierarchy
– frames
• shared characteristics, IS-A relationships
Expert System Successes
•
•
•
•
XCON - configures systems for DEC
Prospector - an mining expert
MYCIN - infectious blood diseases
EMYCIN - Empty MYCIN
Elements of Expert System Shell
• Knowledge Base
– rules
• Working Memory
– facts of current case
• Inference Engine
– applies rules using current set of facts
• Explanation Facility
• CLIPS
Neural Networks & The Brain
• Base on architecture of human brain
–
–
–
–
–
Neurons connected by axons & dendrites
100 billion neurons
1,000 dendrites per neuron
100,000 billion synapses
10 million billion interconnectons per second
How a Neuron Works
Sending
impulses
to next
level of
neurons.
Impulses
come from
other neurons.
When sum of
inputs reaches
a threshold,
neuron fires.
An Artificial Neural Network
w
w
w
w
w
w
Inputs
Hidden
Output
Neural Networks, NN
• NNs learn by using a training set and
adjusting the weights on each connection.
• NNs do not have to be “told” explicit
relationship rules.
• NNs can work with partial inputs.
• NNs cannot explain their results.
• NNs can take a long time to train.
• A NN demonstration