Zero-bias autoencoders and the benefits of co-adapting features.

Roland Memisevic Kishore Reddy Konda David Krueger

Published in: ICLR (Poster) (2015)

Keyphrases

rich set
feature extraction
feature vectors
low level
classification accuracy
denoising
co occurrence
data sets
high level
feature space
image features
benchmark datasets
false positives
svm classifier
key features
invariant features