Abstract

Convolutional neural networks are often used as a model of the primate visual system. However, it is often overlooked how a task that the network performs and statistics of the training set affect the representation of information in the latent space of the model. This study demonstrates that the properties of artificial neurons in the first two convolutional layers represent the signal statistics (correlation coefficients R=0.63 and R=0.44), whereas the similarity between the space of the problem and information encoding in hidden layers gradually increases in the final convolutional layers (R=0.35), reaching a value of 0.73 in the fully-connected layers. At the final stages of the processing, a category is encoded using a unique set of features, characterized by no or little overlapping with other categories. Thus, in order to increase similarity between the visual system and its model, it is important to maintain a training set and a problem space of the model coherent to those of a biological organism.

© 2021 Optical Society of America

Processing of chromatic information in a deep convolutional neural network

Alban Flachot and Karl R. Gegenfurtner
J. Opt. Soc. Am. A 35(4) B334-B346 (2018)

Wavefront reconstruction with artificial neural networks

Hong Guo, Nina Korablinova, Qiushi Ren, and Josef Bille
Opt. Express 14(14) 6456-6462 (2006)

Turbulence correction with artificial neural networks

Sanjaya Lohani and Ryan T. Glasser
Opt. Lett. 43(11) 2611-2614 (2018)

Artificial neural network for the determination of Hubble Space Telescope aberration from stellar images

Todd K. Barrett and David G. Sandler
Appl. Opt. 32(10) 1720-1727 (1993)

Diopter measurement based on an artificial neural network

Tao Jin and Xuan Gao
Appl. Opt. 60(34) 10671-10679 (2021)