Adaptive Activation Function Generation for Artificial Neural Networks through Fuzzy Inference with Application in Grooming Text Categorisation

Zheming Zuo, Jie Li, Bo Wei, Longzhi Yang, Fei Chao, Nitin Naik

Research output: Chapter in Book/Published conference outputConference publication

Abstract

The activation function is introduced to determine the output of neural networks by mapping the resulting values of neurons into a specific range. The activation functions often suffer from 'gradient vanishing', 'non zero-centred function outputs', 'exploding gradients', and 'dead neurons', which may lead to deterioration in the classification performance. This paper proposes an activation function generation approach using the Takagi-Sugeno-Kang inference in an effort to address such challenges. In addition, the proposed method further optimises the coefficients in the activation function using the genetic algorithm such that the activation function can adapt to different applications. This approach has been applied to a digital forensics application of online grooming detection. The evaluations confirm the superiority of the proposed activation function for online grooming detection using an unbalanced data set.

Original languageEnglish
Title of host publication2019 IEEE International Conference on Fuzzy Systems, FUZZ 2019
PublisherIEEE
ISBN (Electronic)9781538617281
DOIs
Publication statusPublished - 10 Oct 2019
Event2019 IEEE International Conference on Fuzzy Systems, FUZZ 2019 - New Orleans, United States
Duration: 23 Jun 201926 Jun 2019

Publication series

NameIEEE International Conference on Fuzzy Systems
Volume2019-June
ISSN (Print)1098-7584

Conference

Conference2019 IEEE International Conference on Fuzzy Systems, FUZZ 2019
Country/TerritoryUnited States
CityNew Orleans
Period23/06/1926/06/19

Fingerprint

Dive into the research topics of 'Adaptive Activation Function Generation for Artificial Neural Networks through Fuzzy Inference with Application in Grooming Text Categorisation'. Together they form a unique fingerprint.

Cite this