Date of Graduation

12-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Computer Science (PhD)

Degree Level

Graduate

Department

Computer Science & Computer Engineering

Advisor/Mentor

Gauch, Susan

Committee Member

Rajagopala, Abh

Second Committee Member

Zhang, Lu

Third Committee Member

Zhang, Qingyang

Keywords

artificial intelligence; emotion analysis; natural language processing

Abstract

Emotion analysis is a branch of artificial intelligence and natural language processing focused on recognizing emotions hidden throughout various forms of digital data, including text, images, and multi-modal representations. In this dissertation, we present four published and planned works that investigate different methodologies for natural language analysis tasks using deep learning techniques. The first published work we present considers the task of identifying fake news using various text and emotion representations. We demonstrate that emotion representations combined with word embedding techniques can improve the accuracy of fake news classification. Our second published work further investigates the fake news classification task by leveraging deep neural networks and neural language models. In this work, we propose several deep neural network architectures and evaluate the performance of models using current state-of-the-art neural language model representations. We demonstrate that emotion representations can improve the accuracy of fake news classification. In our third work that is completed and ready for submission, rather than relying on lexicon-based methods of emotion vector formation, we construct a deep neural network model that learns the latent emotion representations in text by aligning emotion vectors with neural language models for a multi-label emotion classification task. Lastly, our fourth planned publication considers the task of aligning emotion vectors with multi-modal emotion representations for joint image and text representations. This allows us to encode image and text with representations into a multidimensional space that can be used for the purposes of classification. As a result, our multimodal model utilizing joint visual linguistic features combined with emotion analysis is expected to further improve the task of fake news identification in a multimodel context.

Download

Share

COinS