KMi Stadium

Understanding Misogynoir Online: Challenges in Identifying Intersectional Hate Speech

Dr. Joseph Kwarteng is a Research Associate at the Knowledge Media Institute (KMi) of the Open University.

In today's digital age, where social networking sites serve as a nexus for diverse cultures and identities worldwide, we witness the dark side of this connectivity through the rise of hate speech. This form of digital aggression, targeting specific genders and ethnicities, proliferates at an alarming rate, transcending geographical, social, and political barriers, thereby amplifying its detrimental impact. The challenge of moderating online spaces to safeguard vulnerable communities has never been more pressing. In this seminar, we delve into the phenomenon of "Misogynoir", a distinct category of intersectional hate speech that compounds racial and gender-based biases against Black women and girls. We will offer insights from our comprehensive investigation into misogynoir, aiming to shed light on the complexities and challenges associated with effectively understanding and identifying this form of hate speech. Lastly, we will dedicate 5-10 minutes of discussion about the commonly held perceptions and personal experiences related to the PhD Viva. This will include topics such as the overall experience, the emotions it evokes, interactions with the examiners, and the reactions to the process.

KMi Seminar took place on 30th April 2024.

Tracking bias creep in machine learning with covariance analysis

Angel Pavon Perezis a PhD Candidate at the Knowledge Media Institute, The Open University and is researching bias in artificial intelligence models in financial services.

Demand for fairer machine learning models is rapidly growing due to their increasing use in many decision-making processes. Several methods have been developed to detect and mitigate the bias of these models. One common approach for addressing such bias is simply dropping the sensitive attribute from the training data (e.g. gender). Such an approach is limited by focusing on the fairness of the training process rather than on the output model. This kind of simplistic approaches overlook the fact that sensitive attributes can be indirectly represented by other attributes in the data (e.g. maternity leave taken). However, there is currently little research aiming at understanding how covariance in data can contribute to the propagation of bias in machine learning models. In this seminar, we use feature selection techniques and statistical tests to study the covariance of these attributes and show how this covariance can help explain model bias in credit risk data. We further demonstrate how fairness can be significantly improved by eliminating the related attributes and the subsequent impact on model accuracy.