Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

An adaptively trained neural network.

D C Park¹, M A El-Sharkawi, R J Marks

¹Dept. of Electr. and Comput. Eng., Florida Int. Univ., Miami, FL.

IEEE Transactions on Neural Networks

|January 1, 1991

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

Ectopic pregnancy occurring in the remnant tube of a previous adnexectomy: a case report.

Clinical and experimental obstetrics & gynecology·2018

Same author

A Case of Inflamed Strangulated Inguinal Hernia: Operation: Removal of 4½ Inches of Omentum: Recovery.

The Indian medical gazette·2017

Same author

Case of Ovariotomy.

The Indian medical gazette·2017

Same author

Prostatic Enlargement in Natives.

The Indian medical gazette·2017

Same author

A Case of Intestinal Obstruction, Due to a Stone; Operation : Recovery.

The Indian medical gazette·2017

Same author

Bone mineral density among Korean females aged 20-50 years: influence of age at menarche (The Korea National Health and Nutrition Examination Survey 2008-2011).

Osteoporosis international : a journal established as result of cooperation between the European Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA·2017

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 an adaptive training procedure for artificial neural networks to handle changing data. The adaptively trained neural network (ATNN) effectively learns new information without forgetting previous data, outperforming standard methods in electric load forecasting.

Area of Science:

Artificial Intelligence
Machine Learning
Neural Networks

Background:

Traditional neural networks struggle with nonstationary data where data distributions change over time.
Layered perceptron networks require retraining for new data, which can be inefficient and lead to catastrophic forgetting.

Purpose of the Study:

To develop an adaptive training procedure for artificial neural networks that can adjust to slowly varying nonstationary data.
To create an adaptively trained neural network (ATNN) that can incorporate new, potentially conflicting data without compromising performance on older data.
To improve the efficiency and adaptability of neural networks in dynamic environments.

Main Methods:

Proposed a novel training procedure based on nonlinear programming techniques to adapt neural network weights.

Related Experiment Videos

Developed an adaptively trained neural network (ATNN) capable of handling conflicting training data.

Implemented a weighting mechanism to prioritize new data based on its significance.

Main Results:

The ATNN demonstrated the ability to adapt to new training data that conflicted with previous data.
The network's performance on previously learned data remained unaffected by the adaptation to new data.
The adaptive algorithm showed superior performance compared to conventionally trained layered perceptrons in electric load forecasting.

Conclusions:

The proposed adaptive training procedure enables artificial neural networks to effectively learn from nonstationary data.
ATNNs offer a robust solution for dynamic environments where data evolves over time.
This method provides a significant advancement for applications like electric load forecasting, enhancing prediction accuracy and adaptability.