Date of Award

2-14-2023

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Science

First Advisor

Nitin Agarwal

Abstract

Recent technological advancements have pushed humans past the boundaries of a computer screen. They have facilitated human-computer-interaction that were once inconceivable including emotional and audio engagement. Users are now able to utilize technology in more human-like ways and instantaneously transfer depth of opinions and emotions. This global prominence of modern technology, specifically with social media, has spawned a new norm, in which it is now a reasonable expectation of encountering high online toxicity on social media platforms. In addition, the complexities of human emotions make it challenging to determine whether toxic comments trigger certain emotions or vice versa. This study explores and bridges the gap through the characterization of audio to text content production based on content creators’ emotions and toxicity levels utilizing machine learning techniques to detect and cluster features. This thesis also provides at least a partial answer to the following research questions: RQ1: What types of emotions are being pushed out in multimedia contents? RQ2: Can content be identified as toxic? If so, what are the toxicity levels? RQ3: What is the relationship between online emotions and toxicity for social media content? The research described in this thesis answers these questions utilizing a characterization-based assessment of emotions and toxicity utilizing transfer learning techniques from the audio speech recognition framework, Emonet, and a pre-trained machine learning model, Detoxify. The research also explored the toxicity levels relative to emotion scores to identify any identifiable relationships.

Download

Share

COinS