Evolutionary Computation and Machine Learning for Fake News Detection
Metadata Field | Value | Language |
---|---|---|
dc.contributor.advisor | Dozier, Gerry | |
dc.contributor.author | Smith, Marcellus | |
dc.date.accessioned | 2022-05-06T19:13:52Z | |
dc.date.available | 2022-05-06T19:13:52Z | |
dc.date.issued | 2022-05-06 | |
dc.identifier.uri | https://etd.auburn.edu//handle/10415/8234 | |
dc.description.abstract | Misinformation or fake news has a history of being weaponized in order to deceive or mislead a target audience. Presently, the ease of generation, dissemination, and effects that are achievable by leveraging fake news makes fake news detection a critical issue that needs to be addressed. Effectively detecting fake news is faced with many challenges. Some of these challenges include the countless features that are able to be extracted from a text corpus, adversarial effects on fake news detection systems, and wide scale propagation of information. This document outlines research aimed at tackling and overcoming these challenges. Genetic and evolutionary feature selection (GEFeS) is combined with fake news detection to improve detection accuracy and identify essential features; this work was published in IEEE Symposium Series on Computational Intelligence [1]. A novel method is proposed for selecting subsets adversarial examples for adversarial training in order to strengthen the defensive posture of a machine learning system; this work was published in IEEE Congress on Evolutionary Computation [2]. Determining the propagation of information types during an infodemic and building an auto-labeler; this work was published in IEEE SoutheastCon. The novel method, genetic adversarial training, is extended to overcome learning constraints and to incorporate multi-objective optimization; this work was published in IEEE Symposium Series on Computational Intelligence [3]. The work concerning the propagation of information types is extended to examine a broader scope of information and leverages a more advanced auto-labeling method; this work is pending publication. | en_US |
dc.rights | EMBARGO_GLOBAL | en_US |
dc.subject | Computer Science and Software Engineering | en_US |
dc.title | Evolutionary Computation and Machine Learning for Fake News Detection | en_US |
dc.type | PhD Dissertation | en_US |
dc.embargo.length | MONTHS_WITHHELD:36 | en_US |
dc.embargo.status | EMBARGOED | en_US |
dc.embargo.enddate | 2025-05-06 | en_US |