Similarities of error regularization, sigmoid gain scaling, target smoothing, and training with jitter.
Russell ReedRobert J. Marks IISeho OhPublished in: IEEE Trans. Neural Networks (1995)
Keyphrases
- training error
- early stopping
- training set
- bilateral filtering
- similarity measure
- total variation regularization
- stochastic gradient descent
- training phase
- hidden layer
- training process
- belief networks
- error rate
- supervised learning
- training examples
- denoising
- pairwise
- smoothing algorithm
- bayesian networks
- image processing
- smoothing parameter