Show Reference: "Self-Organization of Orientation Maps, Lateral Connections, and Dynamic Receptive Fields in the Primary Visual Cortex"

Self-Organization of Orientation Maps, Lateral Connections, and Dynamic Receptive Fields in the Primary Visual Cortex In: International Conference on Artificial Neural Networks — ICANN 2001, Vol. 2130 (2001), pp. 1147-1152, doi:10.1007/3-540-44668-0_160 by Cornelius Weber edited by Georg Dorffner, Horst Bischof, Kurt Hornik
@inproceedings{weber-2001,
    abstract = {We set up a combined model of sparse coding bottom-up feature detectors and a subsequent attractor with horizontal weights. It is trained with filtered grey-scale natural images. We find the following results on the connectivity: (i) the bottom-up connections establish a topographic map where orientation and frequency are represented in an ordered fashion, but phase randomly. (ii) the lateral connections display local excitation and surround inhibition in the feature spaces of position, orientation and frequency. The results on the attractor activations after an interrupted relaxation of the attractor cells as a response to a stimulus are: (i) Contrast-response curves measured as responses to sine gratings increase sharply at low contrasts, but decrease at higher contrasts (as reported for cells which are adapted to low contrasts [1]). (ii) Orientation tuning curves of the attractor cells are more peaked than those of the feature cells. They have reasonable contrast invariant tuning widths, however, the regime of gain (along the contrast axis) is small before saturation is reached. (iii) The optimal response is roughly phase invariant, if the attractor is trained to predict its input when images move slightly.},
    author = {Weber, Cornelius},
    booktitle = {International Conference on Artificial Neural Networks — ICANN 2001},
    citeulike-article-id = {5373040},
    citeulike-linkout-0 = {http://dx.doi.org/10.1007/3-540-44668-0\_160},
    citeulike-linkout-1 = {http://www.springerlink.com/content/3ah800j45r15949h},
    citeulike-linkout-2 = {http://link.springer.com/chapter/10.1007/3-540-44668-0\_160},
    doi = {10.1007/3-540-44668-0\_160},
    editor = {Dorffner, Georg and Bischof, Horst and Hornik, Kurt},
    journal = {Artificial Neural Networks — ICANN 2001},
    keywords = {ann, learning, self-organization, unsupervised-learning},
    pages = {1147--1152},
    posted-at = {2015-03-05 10:18:26},
    priority = {2},
    publisher = {Springer Berlin Heidelberg},
    series = {Lecture Notes in Computer Science},
    title = {{Self-Organization} of Orientation Maps, Lateral Connections, and Dynamic Receptive Fields in the Primary Visual Cortex},
    url = {http://dx.doi.org/10.1007/3-540-44668-0\_160},
    volume = {2130},
    year = {2001}
}

See the CiteULike entry for more info, PDF links, BibTex etc.

Weber presents a Helmholtz machine extended by adaptive lateral connections between units and a topological interpretation of the network. A Gaussian prior over the population response (a prior favoring co-activation of close-by units) and training with natural images lead to spatial self-organization and feature-selectivity similar to that in cells in early visual cortex.

Weber presents a continuous Hopfield-like RNN as a model of complex cells in V1. This model receives input from a sparse coding generative Helmholtz machine, described earlier as a model of simple cells in V1, and which produces topography by coactivating neighbors in its "sleep phase". The complex cell model with its horizontal connections is trained to predict the simple cells' activations, while input images undergo small random shifts. The trained network features realistic centre-surround weight profiles (in position- and orientation-space) and sharpened orientation tuning curves.