Jove
Visualize
Contact Us
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

Related Experiment Videos

Network Synthesis through Data-Driven Growth and Decay.

Demetri Psaltis1, Chuanyi Ji

  • 1Department of Electrical Engineering, California Institute of Technology, Troy, NY 12180-3590, USA

Neural Networks : the Official Journal of the International Neural Network Society
|August 1, 1997
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

Training of physical neural networks.

Nature·2025
Same author

Training hybrid neural networks with multimode optical nonlinearities using digital twins.

Nanophotonics (Berlin, Germany)·2025
Same author

Guided nonlinear optics for information processing.

Nanophotonics (Berlin, Germany)·2025
Same author

Resource-efficient photonic networks for next-generation AI computing.

Light, science & applications·2025
Same author

From 3D to 2D and back again.

Nanophotonics (Berlin, Germany)·2024
Same author

Learning to image and compute with multimode optical fibers.

Nanophotonics (Berlin, Germany)·2024
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
Same journal

IGFD-Net: Illumination-guided frequency decoupling for polarization image fusion.

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

Multiple-Strategies dung beetle optimizer and its applications in engineering optimization and bankruptcy prediction.

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

Aggregating global-scale pixel-wise forgery cues within a graph.

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

Finite-Time intermittent control for secure synchronization of Neutral-Type stochastic delayed neural networks under aperiodic DoS attacks.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

This study introduces an algorithm for adding and deleting (ADDEL) network resources to create small, well-structured neural networks. The ADDEL algorithm effectively optimizes network size for better generalization performance.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Neural Networks

Background:

  • Developing efficient neural network architectures is crucial for effective machine learning.
  • Finding optimal network structures that balance complexity and generalization remains a challenge.

Purpose of the Study:

  • To develop an algorithm (ADDEL) for optimizing feed-forward multilayer network size and structure.
  • To achieve minimal network size while maintaining appropriate structural complexity for learning.

Main Methods:

  • The ADDEL algorithm alternates between adding and deleting network resources (connections, units, layers).
  • A sensitivity measure and probability rule guide resource addition and deletion to minimize output error.
  • Generalization error from a validation set controls learning phases and termination.

Related Experiment Videos

Main Results:

  • Simulations, including handwritten digit recognition, show the algorithm's effectiveness.
  • The ADDEL algorithm successfully identifies appropriate structures for small, generalizable networks.
  • The study investigates the impact of network size on generalization capabilities.

Conclusions:

  • The ADDEL algorithm provides an effective method for designing compact and efficient neural networks.
  • Optimizing network structure through ADDEL leads to improved generalization performance.
  • Network size plays a significant role in achieving good generalization.