Presentation is loading. Please wait.

Presentation is loading. Please wait.

Introduction to NeuroFuzzy Technologies © INFORM 1990-1998Slide 1 Tutorial and Workshop © Constantin von Altrock Inform Software Corporation 2001 Midwest.

Similar presentations


Presentation on theme: "Introduction to NeuroFuzzy Technologies © INFORM 1990-1998Slide 1 Tutorial and Workshop © Constantin von Altrock Inform Software Corporation 2001 Midwest."— Presentation transcript:

1 Introduction to NeuroFuzzy Technologies © INFORM 1990-1998Slide 1 Tutorial and Workshop © Constantin von Altrock Inform Software Corporation 2001 Midwest Rd. Oak Brook, IL 60521, U.S.A. German Version Available! Phone 630-268-7550 Fax 630-268-7554 Email: fuzzy@informusa.com Internet: www.fuzzytech.com Tutorial and Workshop © Constantin von Altrock Inform Software Corporation 2001 Midwest Rd. Oak Brook, IL 60521, U.S.A. German Version Available! Phone 630-268-7550 Fax 630-268-7554 Email: fuzzy@informusa.com Internet: www.fuzzytech.com Combining Neural Networks and Fuzzy Logic X XNeural Net Basics X XTraining Neural Nets X XCombining Neural and Fuzzy X XTraining Fuzzy Logic Systems X XConvergence of Technologies X XExamples Combining Neural Networks and Fuzzy Logic X XNeural Net Basics X XTraining Neural Nets X XCombining Neural and Fuzzy X XTraining Fuzzy Logic Systems X XConvergence of Technologies X XExamples

2 Neural Net Basics: - Neuron Model - Neural Net Basics: - Neuron Model - © INFORM 1990-1998Slide 2 X XMultiple Inputs, One Output X XThe Output Signal is the Activation Level of the Neuron X XThe Inputs Stem From Outputs of Other Neurons X XInputs Wired to the Neuron Using Exciting Synapses Increase Activation Level X XInputs Wired to the Neuron Using Inhibiting Synapses Decrease Activation Level X XMultiple Inputs, One Output X XThe Output Signal is the Activation Level of the Neuron X XThe Inputs Stem From Outputs of Other Neurons X XInputs Wired to the Neuron Using Exciting Synapses Increase Activation Level X XInputs Wired to the Neuron Using Inhibiting Synapses Decrease Activation Level WARNING: This Neuron Model is a Strong Simplification of “Mother Nature”

3 Neural Net Basics: - Mathematical Model - Neural Net Basics: - Mathematical Model - © INFORM 1990-1998Slide 3

4 Neural Net Basics: - Multilayer Nets - Neural Net Basics: - Multilayer Nets - © INFORM 1990-1998Slide 4

5 Training Neural Nets: - Pavlovs’ Dogs - Training Neural Nets: - Pavlovs’ Dogs - © INFORM 1990-1998Slide 5 Hebb’s Learning Rule: Increase weight to active input neuron, if the output of this neuron should be active,decrease weight to active input neuron, if the output of this neuron should be inactive.

6 Combining Neural and Fuzzy Combining Neural and Fuzzy © INFORM 1990-1998Slide 6 X XNeural Networks have their Strengths X XFuzzy Logic has its Strengths Neural Nets Knowledge Representation Fuzzy Logic Trainability Implicit, the system cannot be easy interpreted or modified (-) Trains itself by learning from data sets (+++) Explicit, verification and optimization easy and efficient (+++) None, you have to define everything explicitly (-) Get “best of both worlds”: Explicit Knowledge Representation from Fuzzy Logic with Training Algorithms from Neural Nets

7 Training Fuzzy Logic Systems Training Fuzzy Logic Systems © INFORM 1990-1998Slide 7 X XMany Different Ways Exist to Train a Fuzzy Logic System X XNeuroFuzzy := Use Error Backpropagation X XEmulate Fuzzy Logic System as Neural Net X XEach Component of a Fuzzy Logic System is Represented as Part of a Neural Net X XApply EPG to this ‘Neural Net’ X X  : EPG Requires Differentiability X X  : Use Gradient Estimators X X  : Use Fuzzy Associative Memories X XMany Different Ways Exist to Train a Fuzzy Logic System X XNeuroFuzzy := Use Error Backpropagation X XEmulate Fuzzy Logic System as Neural Net X XEach Component of a Fuzzy Logic System is Represented as Part of a Neural Net X XApply EPG to this ‘Neural Net’ X X  : EPG Requires Differentiability X X  : Use Gradient Estimators X X  : Use Fuzzy Associative Memories

8 Convergence of Technologies © INFORM 1990-1998Slide 8 Year: 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 Year: 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 Computing: Relay/Valve Based Transistors Small Scale Integration Large Scale Integration Artificial Intelligence Computing: Relay/Valve Based Transistors Small Scale Integration Large Scale Integration Artificial Intelligence Neural Networks: Neuron Model (McCulloch/Pitts) Training Rules (Hepp) Delta Rule (Wirow/Hoff) Multilayer Perceptron, XOR Hopfield Model (Hopfield/Tank) Backpropagation (Rumelhart) Bidir. Assoc. Mem. (Kosko) Neural Networks: Neuron Model (McCulloch/Pitts) Training Rules (Hepp) Delta Rule (Wirow/Hoff) Multilayer Perceptron, XOR Hopfield Model (Hopfield/Tank) Backpropagation (Rumelhart) Bidir. Assoc. Mem. (Kosko) Fuzzy Logic: Seminal Paper (Zadeh) Fuzzy Control (Mamdani) Broad Application in Japan Broad Application in Europe Broad Application in the U.S. Fuzzy Logic: Seminal Paper (Zadeh) Fuzzy Control (Mamdani) Broad Application in Japan Broad Application in Europe Broad Application in the U.S. Soft Computing, NeuroFuzzy


Download ppt "Introduction to NeuroFuzzy Technologies © INFORM 1990-1998Slide 1 Tutorial and Workshop © Constantin von Altrock Inform Software Corporation 2001 Midwest."

Similar presentations


Ads by Google