My main research interests are in studying and developing methods to make machine learning (ML) algorithms more robust to enable their application in safety critical domains. Specifically, I am interested in the verification of ML algorithms, which involves building methods to prove certain properties about the outputs of a model. I chose to do my PhD with the STAI CDT because my research interests naturally align with the aims of promoting safety within AI. Having previously worked in the domain of applying ML to healthcare highlighted a key problem within machine learning – a lack of trust in domain experts, which stems from a lack of explainability as well as observed fragility in predictions, which is difficult to understand, quantify, and subsequently mitigate.
The STAI CDT has provided me with an excellent platform to pursue my research interests as well as interact with like-minded peers. Furthermore, the additional training provided by the CDT has proven to be invaluable in terms of the additional perspective and context that it has provided me for my research.
–
Masters Qualification: MEng. Aeronautical Engineering, Imperial College London
LinkedIn: https://www.linkedin.com/in/atri-sharma/
