Jove
Visualize
Contact Us

Related Experiment Videos

Linear recursive distributed representations.

Thomas Voegtlin1, Peter F Dominey

  • 1INRIA, Campus Scientifique, B.P. 239, F-54506 Vandoeuvre-Les-Nancy Cedex, France. voegtlin@loria.fr

Neural Networks : the Official Journal of the International Neural Network Society
|June 7, 2005
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Synthesizing the temporal self: robotic models of episodic and autobiographical memory.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2024
Same author

Explaining the neural activity distribution associated with discrete movement sequences: Evidence for parallel functional systems.

Cognitive, affective & behavioral neuroscience·2018
Same author

Narrative Constructions for the Organization of Self Experience: Proof of Concept via Embodied Robotics.

Frontiers in psychology·2017
Same author

Editorial: The Temporal Dynamics of Cognitive Processing.

Frontiers in psychology·2016
Same author

Behavioral Regulation and the Modulation of Information Coding in the Lateral Prefrontal and Cingulate Cortex.

Cerebral cortex (New York, N.Y. : 1991)·2014
Same author

Discrete sequence production with and without a pause: the role of cortex, basal ganglia, and cerebellum.

Frontiers in human neuroscience·2013
Same journal

Dynamic analysis and reliable mechanical optimization application of ring HNN effected with a memristive neuron.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

DAFF-Net: A detection and search method for small-scale low surface brightness galaxies.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Quasi-synchronization for complex networks with hybrid pinning intermittent control.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Physics-encoded convolutional neural operators for parametric PDEs: A convergence-guaranteed framework via pre-computed kernel fields.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

This study introduces a modified Recursive Auto-Associative Memory (RAAM) network that overcomes limitations in representing complex structures. The new model demonstrates superior generalization capabilities for structured data, enhancing artificial intelligence development.

Area of Science:

  • Artificial Intelligence
  • Cognitive Science
  • Computational Neuroscience

Background:

  • Connectionist networks struggle with systematicity, limiting generalization to novel combinations of learned structures.
  • Existing Recursive Auto-Associative Memory (RAAM) models face challenges in representing and generalizing complex, structured data.
  • The inability to generalize systematically is a key criticism of current neural network architectures.

Purpose of the Study:

  • To address the systematicity criticism of connectionist networks by modifying Recursive Auto-Associative Memory (RAAM).
  • To develop a network capable of representing and generalizing novel combinations of constituents within complex structures.
  • To enhance the capacity and generalization abilities of Recursive Auto-Associative Memory (RAAM) models.

Main Methods:

Related Experiment Videos

  • Modification of Pollack's Recursive Auto-Associative Memory (RAAM) using linear units.
  • Training the network with Oja's rule, a generalization of Principal Component Analysis (PCA) for tree-structured data.
  • Employing linear combinations of learned representations to construct new complex structures.

Main Results:

  • Achieved unprecedented generalization capabilities for structured data.
  • Demonstrated a capacity orders of magnitude higher than standard RAAM trained with back-propagation.
  • Preserved regularities from the training set in newly formed representations, showing developmental effects.

Conclusions:

  • The modified RAAM network effectively addresses the systematicity limitations of connectionist models.
  • This approach offers significant improvements in generalization and capacity for structured data representation.
  • The findings suggest potential parallels with human cognitive development in learning and generalization.