Paper Info
Title
Diversity priors for learning early visual features.
Abstract
This paper investigates how utilizing diversity priors can discover early visual features that resemble their biological counterparts. The study is mainly motivated by the sparsity and selectivity of activations of visual neurons in area Vi. Most previous work on computational modeling emphasizes selectivity or sparsity independently. However, we argue that selectivity and sparsity are just two epiphenomena of the diversity of receptive fields, which has been rarely exploited in learning. In this paper, to verify our hypothesis, restricted Boltzmann machines (RBMs) are employed to learn early visual features by modeling the statistics of natural images. Considering RBMs as neural networks, the receptive fields of neurons are formed by the inter-weights between hidden and visible nodes. Due to the conditional independence in RBMs, there is no mechanism to coordinate the activations of individual neurons or the whole population. A diversity prior is introduced in this paper for training RBMs. We find that the diversity prior indeed can assure simultaneously sparsity and selectivity of neuron activations. The learned receptive fields yield a high degree of biological similarity in comparison to physiological data. Also, corresponding visual features display a good generative capability in image reconstruction.
Year
DOI
Venue
2015
10.3389/fncom.2015.00104
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE
Keywords
Field
DocType
Markov networks,V1 simple cell,diversity prior,inhibition,restricted Boltzmann machine
Iterative reconstruction,Receptive field,Population,Restricted Boltzmann machine,Neuroscience,Boltzmann machine,Conditional independence,Computer science,Artificial intelligence,Artificial neural network,Prior probability,Machine learning
Journal
Volume
Citations 
PageRank 
9
3
0.43
References 
Authors
17
4
Name
Order
Citations
PageRank
Hanchen Xiong1144.40
antonio j rodriguezsanchez230.43
Sandor Szedmak342636.31
Justus Piater450032.76