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

Constructive feedforward neural networks using hermite polynomial activation functions.

Liying Ma1, K Khorasani

  • 1Department of Electrical and Computer Engineering, Concordia University, Montreal, QC, H3G 1M8 Canada.

IEEE Transactions on Neural Networks
|August 27, 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

Natural Variation and Expression of OsCFF1 Affect the Formation of Rice Chalkiness.

Rice (New York, N.Y.)·2026
Same author

Azvudine (FNC), a next-generation cytidine analog with broad inhibitory activity against drug-resistant HIV-1 strains via stabilization of the reverse transcriptase catalytic complex.

Bioorganic chemistry·2026
Same author

Aluminum-Triggered Dealloying for Hierarchical Porous High-Entropy Alloy Electrodes Enabling Industrially Stable Oxygen Evolution.

ACS nano·2026
Same author

Biomembrane-coated Nanoparticles Targeting circHIF1α Suppress Ovarian Cancer Metastasis and Cisplatin Resistance by Mediating System Xc⁻ Inactivation via SLC7A11/SLC3A2 to Induce Ferroptosis in Cancer Stem Cells.

International journal of biological sciences·2026
Same author

YBX1 downregulation links to impaired spermatogonial stem cells function potentially through inhibiting miRNAs production.

iScience·2026
Same author

Hierarchical Porous Collectors with Lithiophilicity and Conductivity Gradients for Regulated Lithium Deposition and Stable Lithium Metal Anodes.

Small (Weinheim an der Bergstrasse, Germany)·2026
Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013
Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011
Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011
Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011
Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011
Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011
See all related articles

This study introduces a novel constructive neural network using unique polynomial activation functions for each neuron. This approach enhances performance compared to traditional networks with identical sigmoidal activation functions.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Neural Networks

Background:

  • Traditional neural networks often use identical activation functions across all hidden units.
  • This uniformity can limit the network's ability to effectively model complex input-output relationships.

Purpose of the Study:

  • To introduce a constructive neural network architecture with diverse activation functions.
  • To enhance the network's capacity for capturing intricate data patterns.

Main Methods:

  • A one-hidden-layer constructive network is proposed.
  • Each hidden unit utilizes a unique polynomial activation function, specifically orthonormal Hermite polynomials.
  • Both structural and functional level adaptation methodologies are employed during network construction.

Related Experiment Videos

Main Results:

  • The proposed network demonstrated superior performance in simulations.
  • Compared networks with identical sigmoidal activation functions, the new architecture showed significant improvements.
  • The use of diverse polynomial activation functions led to more effective input-output mapping.

Conclusions:

  • Constructive neural networks with varied polynomial activation functions offer improved performance.
  • Functional level adaptation with unique neuron activation functions enhances modeling capabilities.
  • The proposed approach provides a more effective alternative to standard neural network designs.