Recurrent Dendritic Neural Network

Jiayi Li, Zhipeng Liu, Yaotong Song, Shangce Gao*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In the realm of deep learning, the prevalent use of the simplistic McCulloch-Pitts Neuron as a foundational building block lacks the interpretability required for complex models. In contrast, the dendritic neuron model (DNM) closely emulates biological neural networks and has demonstrated its effectiveness in diverse classification and prediction tasks. This paper introduces the Recurrent Dendritic Neural Network (RDNN), a novel artificial neural network based on the DNM framework. RDNN offers substantial advantages over traditional neural networks, particularly in financial time series forecasting. We propose that the incorporation of DNM as the fundamental unit in artificial neural networks enhances computational capabilities, positioning it at the forefront of the next generation of deep learning technologies.

Original languageEnglish
Title of host publicationIEEE ITAIC 2023 - IEEE 11th Joint International Information Technology and Artificial Intelligence Conference
EditorsBing Xu, Kefen Mou
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1285-1289
Number of pages5
ISBN (Electronic)9798350333664
DOIs
StatePublished - 2023
Event11th Joint International Information Technology and Artificial Intelligence Conference, ITAIC 2023 - Chongqing, China
Duration: 2023/12/082023/12/10

Publication series

NameIEEE Joint International Information Technology and Artificial Intelligence Conference (ITAIC)
ISSN (Print)2693-2865

Conference

Conference11th Joint International Information Technology and Artificial Intelligence Conference, ITAIC 2023
Country/TerritoryChina
CityChongqing
Period2023/12/082023/12/10

Keywords

  • Dendritic Neural Network
  • Recurrent Neural Network
  • Time Series Forecasting

ASJC Scopus subject areas

  • Artificial Intelligence
  • Information Systems

Fingerprint

Dive into the research topics of 'Recurrent Dendritic Neural Network'. Together they form a unique fingerprint.

Cite this