Introduction
An event series for Turing-Roche partnership updates, knowledge sharing and new perspectives. Find out more about the series.
This event will explore the theme of fairness in AI, with relation to health.
About the event
We will be hearing from Leandra Bräuninger, a PhD student at University College London and Adarsh Srivastava, Head of Data and Analytics Quality Assurance at Roche.
Leandra will introduce the concept of fairness for health AI with three pillars: fairness definition, fairness metrics and unfairness mitigation. In the talk, Leandra will describe these pillars conceptually and show a use-case of recent work with the Turing-Roche partnership as illustration. The work slots into the metrics pillar and uses conformal prediction as a personalised uncertainty metric for breast cancer transcriptomic score OncotypeDX, based on a publicly available dataset.
Adarsh will discuss the recently published Roche AI Ethics principles, with particular focus on the aspect of fairness. He will emphasize the pivotal role fairness plays in the development and implementation of AI systems. He will underscore the necessity of transparently addressing any inherent biases in AI models and actively striving to rectify them. Moreover, he will stress the importance of accountability, emphasizing the responsibility of developers and stakeholders to ensure that AI technologies are wielded to promote fairness and societal well-being.
Watch now
You can watch a recording of this event here.
