Presented By: Jasdeep Singh

Models similar to that of the brain and nervous system. Highly parallel
Process information much more like the brain than a serial computer.

Knowledge in neural networks is stored as synaptic weights between neurons.

A Neural Network is an interconnected assembly of simple processing elements, units or nodes, whose functionality is loosely based on the animal neuron. The neuron collects its real inputs from the outputs of neighboring neurons.
W1 Inputs W2 Wn
Processing element

Output

A hybrid system that combines a neural network and a rule-based expert system is called a neural expert system (or a connectionist expert system). we combine advantages of expert systems and neural networks to create a more powerful and effective expert system.

Expert systems rely on logical inferences and decision trees and focus on modelling human reasoning. Neural networks rely on parallel data processing and focus on modelling a human brain. Knowledge in a rule-based expert system is represented by IF-THEN production rules. Knowledge in neural networks is stored as synaptic weights between neurons.
5

Training Data

Neural Knowledge Base New Data

Rule Extract ion R ule: I F - THE N

I n ference E n gine

Explanation Facilities

User Interface

User
6

The heart of a neural expert system is the inference engine. It controls the information flow in the system and initiates inference over the neural knowledge base. A neural inference engine approximate reasoning. also ensures

In a rule-based expert system, the inference engine compares the condition part of each rule with data given in the database. When the IF part of the rule matches the data in the database, the rule is fired and its THEN part is executed. The precise matching is required (inference engine cannot cope with noisy or incomplete data). Neural expert systems use a trained neural network in place of the knowledge base. The input data does not have to precisely match the data that was used in network training. This ability is called approximate reasoning.
8

Neurons in the network are connected by links, each of which has a numerical weight attached to it. The weights in a trained neural network determine the strength or importance of the associated neuron inputs. Output of one neuron is calculated by defining a function upon inputs and their associated weights.

Wings
+1 -0.8

Rule 1
1.0

Bird
+1

Tail
0

-1.6 -0.7 -0.2

Beak
+1

-1.1 2.2

-0.1

Rule 2
0.0 1.0

Plane
1

-1.0

Feathers
+1

2.8 -2.9

-1.6

Rule 3
1.0

Glider
1

Engine
1

-1.1 1.9 -1.3

10

+1 for True, -1 for False, 0 for Unknown Using these we can give semantic interpretation for the activation of any output neuron. For example, if the object has Wings (+1), Beak (+1) and Feathers (+1), but does not have Engine (-1), then we can conclude that this object is Bird (+1):
X Rule1 1 (0.8) 0 (0.2) 1 2.2 1 2.8 (1) (1.1) 5.3 0

YRule 1 YBird 1
11

We can similarly conclude that this object is not Plane:

X Rule 2 1 (0.7) 0 ( 0.1) 1 0.0 1 ( 1.6) ( 1) 1.9 4.2 0

YRule 2 YPlane 1

and not Glider:

X Rule 3 1 (0.6) 0 (1.1) 1 (1.0) 1 (2.9) (1) (1.3) 4.2 0

YRule 3 YGlider 1

12

By attaching a corresponding question to each input neuron, we can enable the system to prompt the user for initial values of the input variables:

Neuron: Wings Neuron: Tail

Question: Does the object have wings?

Neuron: Beak

Question: Does the object have a tail?

Question: Does the object have a beak? Question: Does the object have feathers? Question: Does the object have an engine?

Neuron: Feathers

Neuron: Engine

13

An inference can be made if the known net weighted input to a neuron is greater than the sum of the absolute values of the weights of the unknown inputs.

i 1

xi wi w j
j 1

Xi : input

W : Weight

where i belongs to known, j doesnt belongs to known and n is the number of neuron inputs.

14

Rule 1: IF a1 AND a3 THEN b1 (0.8) Rule 2: IF a1 AND a4 THEN b1 (0.2) Rule 3: IF a2 AND a5 THEN b2 (-0.1 ) Rule 4: IF a3 AND a4 THEN b3 (0.9)
Input Layer Conjunction Layer
1.0 1.0

Rule 5: IF a5 THEN b3 (0.6) Rule 6: IF b1 AND b3 THEN c1 (0.7) Rule 7: IF b2 THEN c1 (0.1) Rule 8: IF b2 AND b3 THEN c2 (0.9)
Disjunction Layer Conjunction Layer Disjunction Layer

a1

R1

0.8

a2

1.0 1.0

R2

0.2

b1

1.0

R6

0.7

a3

1.0 1.0 1.0 1.0

R3

-0.1

b2

1.0 1.0 1.0 1.0

0.1

c1

R7
0.9

c2

a4

R4

0.9

b3

R8

a5

1.0

R5

0.6
15

CBR

KB

NN

GA

FL

Know. rep. Uncertainty Approximation (noisy incomplete data) Adaptable Learnable Interpretable Learns from scratch

3 1 1

4 1 1

1 4 4

2 4 4

4 4 4

4 3 3 4

2 1 4 1

4 4 1 4

4 4 2 1

2 2 4 3

CBR: case based reasoning systems, KB: Knowledge Based system NN: Neural Networks, GA: Genetic Algorithms, FS: Fuzzy Systems 1 representing low and 4 representing a high utility
16

17