Format

Send to:

Choose Destination
See comment in PubMed Commons below
IEEE Trans Neural Netw. 2008 Jan;19(1):168-82. doi: 10.1109/TNN.2007.905852.

Unsupervised segmentation with dynamical units.

Author information

  • 1T.J Watson IBM Research Center, Yorktown Heights, NY 10598, USA. ravirao@us.ibm.com

Abstract

In this paper, we present a novel network to separate mixtures of inputs that have been previously learned. A significant capability of the network is that it segments the components of each input object that most contribute to its classification. The network consists of amplitude-phase units that can synchronize their dynamics, so that separation is determined by the amplitude of units in an output layer, and segmentation by phase similarity between input and output layer units. Learning is unsupervised and based on a Hebbian update, and the architecture is very simple. Moreover, efficient segmentation can be achieved even when there is considerable superposition of the inputs. The network dynamics are derived from an objective function that rewards sparse coding in the generalized amplitude-phase variables. We argue that this objective function can provide a possible formal interpretation of the binding problem and that the implementation of the network architecture and dynamics is biologically plausible.

PMID:
18269948
[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

0 comments
How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for IEEE Engineering in Medicine and Biology Society
    Loading ...
    Write to the Help Desk