Visiting the research group in May Dominik Schindler, Ph.D. candidate in Applied Mathematics at Imperial College London, gave a talk on his latest research on May, 14.
In his talk, titled “Bayesian Knowledge – Situated and Pluriversal Perspectives on Machine Learning”, he explored the fundamental mathematical structures of machine learning (ML) and challenged some of the assumptions underlying ML at a mathematical level. Informed by feminist and decolonial critiques of Western epistemologies, he demonstrated that seemingly innocent mathematical structures can inscribe epistemologies that invalidate other forms of knowledge production, resulting in epistemic violence. Lastly, he explored the potential of subjective Bayesian statistics as an alternative foundation for ML, and examined possible resonances (and dissonances) with situated and pluriversal perspectives.
We thank Dominik Schindler for his enriching and thought-provoking presentation on the mathematical structures of machine learning.