..and episodic memory, cortical self-organization, schema-based vision

[Risto Miikkulainen]

The following papers on
 - the capacity of episodic memory,
 - self-organization in the primary visual cortex, and
 - schema-based scene analysis
are available by anonymous ftp from cs.utexas.edu:pub/neural-nets/papers 
as well. As always, comments are welcome.

-- Risto Miikkulainen

-------------------------------------------------------------------------
moll.convergence-zone.ps.Z (6 pages)

THE CAPACITY OF CONVERGENCE-ZONE EPISODIC MEMORY

Mark Moll(1), Risto Miikkulainen(2), Jonathan Abbey(3)
(1) Department of Computer Science, University of Twente, the Netherlands.
(2) Department of Computer Sciences, The University of Texas at Austin.
(3) Applied Research Laboratories, Austin, TX.

Technical Report AI93-210, December 1993.

Human episodic memory provides a seemingly unlimited storage for
everyday experiences, and a retrieval system that allows us to access
the experiences with partial activation of their components. This paper
presents a computational model of episodic memory inspired by Damasio's
idea of Convergence Zones. The model consists of a layer of perceptual
feature maps and a binding layer. A perceptual feature pattern is coarse
coded in the binding layer, and stored on the weights between layers. A
partial activation of the stored features activates the binding pattern
which in turn reactivates the entire stored pattern.  A worst-case
analysis shows that with realistic-size layers, the memory capacity of
the model is several times larger than the number of units in the model,
and could account for the large capacity of human episodic memory.

-------------------------------------------------------------------------
sirosh.unified.ps.Z (8 pages)

A UNIFIED NEURAL NETWORK MODEL FOR THE SELF-ORGANIZATION OF
TOPOGRAPHIC RECEPTIVE FIELDS AND LATERAL INTERACTION

Joseph Sirosh and Risto Miikkulainen
Department of Computer Sciences, The University of Texas at Austin.

Technical Report AI94-213, January 1994.
			
A self-organizing neural network model for the simultaneous development
of topographic receptive fields and lateral interactions in cortical
maps is presented.  Both afferent and lateral connections adapt by the
same Hebbian mechanism in a purely local and unsupervised learning
process.  Afferent input weights of each neuron self-organize into
hill-shaped profiles, receptive fields organize topographically across
the network, and unique lateral interaction profiles develop for each
neuron.  The resulting self-organized structure remains in a dynamic and
continuously-adapting equilibrium with the input.  The model can be seen
as a generalization of previous self-organizing models of the visual
cortex, and provides a general computational framework for experiments
on receptive field development and cortical plasticity. The model also
serves to point out general limits on activity-dependent
self-organization: when multiple inputs are presented simultaneously,
the receptive field centers need to be initially ordered for stable
self-organization to occur.

[see also sirosh.cooperative-selforganization.tar: "Cooperative Self-
Organization of Afferent and Lateral Connections in Cortical Maps" ]


-------------------------------------------------------------------------
leow.analyzing.ps.Z (11 pages)

ANALYZING SCENES IN A NEURAL NETWORK MODEL OF SCHEMA-BASED VISION

Wee Kheng Leow, Risto Miikkulainen
Department of Computer Sciences, The University of Texas at Austin.

Technical Report AI94-214, February 1994.

A novel approach to object recognition and scene analysis based on
neural network representation of visual schemas is described.  Given an
input scene, the VISOR system focuses attention successively at each
component, and the schema representations cooperate and compete to match
the inputs. The schema hierarchy is learned from examples through
unsupervised adaptation and reinforcement learning.  VISOR learns that
some objects are more important than others in identifying a scene, and
that the importance of spatial relations varies depending on the scene.
It learns three types of visual schemas: (1) rigid spatial layouts of
components used primarily for describing objects; (2) collections of
components located anywhere in the scene for recognizing certain
man-made scenes (such as a dining table); and (3) rough spatial layouts
of regions of uniform texture and no specific shape that are often found
in natural scenes (such as a road scene).  Compared to traditional
rule-based systems, VISOR shows remarkable robustness of recognition,
and is able to indicate the confidence of its analysis as the inputs
differ increasingly from the schemas.  With such properties, VISOR is a
promising first step towards a general vision system that can be used in
different applications after learning the application-specific schemas.

[ see also leow.priming.ps.Z: "Priming, Perceptual Reversal, and
Circular Reaction in A Neural Network Model of Schema-Based Vision" ]