Simplified Fuzzy ARTMAP

Simplified Fuzzy ARTMAP from Soft Computing (Neural Networks)

1. Simplified Fuzzy ARTMAP Presented by : Pradip Sanjay Bankar (BE19F06F004)

2. What is Fuzzy Logic? 🤔 The term “Fuzzy Logic” refers to the logic of approximation. Boolean logic assumes that every fact is either completely true or false. Fuzzy logic allows varying degrees of truth. Computers can apply this logic to represent vague and imprecise ideas like hot, tall, balding, etc.

3. Crisp vs Fuzzy result

4. What is ART Architecture? 🤔 ART refers to Adaptive Resonance Theory . Adaptive - Open to new learning and Resonance - without losing old information. Accepts an input vector and classify it to one of the categories. The category it belongs to is decided by the one the input vector it most resembles to. If the input does not match to any of the existing categories, a new category is created.

5. Simplified ART Module

6. What is Fuzzy ARTMAP? Fuzzy ARTMAP is a neural network architecture for conducting supervised learning in multidimensional setting. When a fuzzy ARTMAP is used in learning model, it’s trained till it classifies all the training data correctly.

7. Fuzzy ARTMAP

8. Let’s understand Fuzzy ARTMAP It incorporates two fuzzy ART Modules ART-a and ART-b linked together via inter-ART Module called MAP field. Map field is used to form predictive associations between categories and to realize the match tracking rule. Whereby, the vigilance parameter of ART-a increases in response to predictive mismatch at ART-b. Match tracking recognizes categories structure so that predictive error is not repeated on subsequent presentations of input.

9. Thank you