Regularization using jittered training data

The authors investigate the training of a layered perceptron with jittered data. They study the effect of generating additional training data by adding noise to the input data and show that is introduces convolutional smoothing of the target function. Training using such jittered data is shown, unde...

Full description

Saved in:
Bibliographic Details
Main Authors: Reed, R., Oh, S., Marks, R.J.
Format: Conference Proceeding
Language:English
Subjects:
Additive noise
Convolution
Distributed computing
Jitter
Lagrangian functions
Multilayer perceptrons
Noise generators
Random variables
Smoothing methods
Training data
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The authors investigate the training of a layered perceptron with jittered data. They study the effect of generating additional training data by adding noise to the input data and show that is introduces convolutional smoothing of the target function. Training using such jittered data is shown, under a small variance assumption, to be equivalent to Lagrangian regularization with a derivative regularizer. Training with jitter allows regularization within the conventional layered perceptron architecture.< >
DOI:10.1109/IJCNN.1992.227178