ME537-NeuralNets-part1

ME537-NeuralNets-part1 - ME 537: Learning-Based Control...

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon
ME 537: Learning-Based Control Kagan Tumer Oregon State University Week 1, Lecture 2 Neural Network Basics Announcements: HW 1 Due on 10/8 Data sets for HW 1 are online Project selection 10/11 Suggested reading : NN survey paper (Zhang) Chap 1, 2 and Sections 4.1 to 4.5 in Passino Learning From Data Kagan Tumer Oregon State University ! ! y = f(x) = a x + b Linear regression (parametric)
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Learning From Data Kagan Tumer Oregon State University ! ! y = f(x) = a x 2 + b x + c Polynomial regression (parametric) ! y = f(x) = a x + b ??? Learning From Data Kagan Tumer Oregon State University ! ! y = f(x) = a n x n + a n-1 x n-1 + … + a 2 x 2 + a 1 x + a 0 Polynomial regression (parametric)
Background image of page 2
Learning From Data Kagan Tumer Oregon State University ! ! y = f(x) = ?? Neural Networks for Nonlinear Control Kagan Tumer Oregon State University ! ! Motivation: ! Control a system with nonlinear dynamics ! Robot ! Satellite ! Air vehicle ! Do we know what the good control strategies are? ! Yes: “teach” neural network those strategies ! Drive a car and record good driver actions for each state ! Fly a helicopter and record good pilot actions for each state ! No: have a neural network discover those strategies ! Let car drive around and provide feedback on performance
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Neural Networks Kagan Tumer Oregon State University ! ! Why Neural Networks? ! Units or Neurons ! Neural Network Architectures ! Activation Functions ! Single Layer Feed Forward Networks ! Multi Layer Feed Forward Networks ! Error Backpropagation ! Implementation Issues Why Neural Networks? Kagan Tumer Oregon State University ! ! Neural Network: A massively parallel distributed processor made up of simple processing units. It stores “knowledge”. ! An artificial neural network is similar to the brain in that: ! Knowledge is acquired by the network from its environment through a learning process ! Interneuron connection strengths (synaptic weights) are used to store the acquired knowledge !
Background image of page 4
Image of page 5
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 10/03/2010 for the course ME 537 taught by Professor Kegantumer during the Fall '10 term at Oregon State.

Page1 / 18

ME537-NeuralNets-part1 - ME 537: Learning-Based Control...

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online