Abstract: Most of today's Ethical AI debate at events revolves around how to do business in an ethical manner (company values and ethics boards), how to build teams to inforce ethical practices (e.g. diversity), and how to work in cross-domain settings (legal/tech/HR/etc.). Although all very important, the core factor of Ethical AI is not nearly getting enough attention.
Applying ethical AI in day to day operations is a highly technical undertaking. Some of the main aspects of Ethical AI (algorithmic fairness and bias, interpretability, robustness, privacy by design) have to be taken into account from the very beginning of the data science process, e.g. when defining a classifier's loss function.
To do so, data scientists need training and tooling. Knowing how to translate legal terms like disparate impact and treatment into technical metrics like group fairness, predictive parity, equalized odds or treatment equality is not something that is touched upon in typical data science courses.
Knowing which metrics to optimize for and how to deal with bias and unfairness when the sensitive attributes are unknown is an open and actively researched problem.
In this talk, we will dive into the technicalities of 'fairness'. We will review how fairness can be defined, measured and enforced with practical examples. This talk is focused on data scientists and aims to provide practical tips and methods to start dealing with fairness and bias in both data and models.
Bio: Vincent serves as Chief Innovation Officer at Sentiance, a scale-up that uses AI to model, predict and coach human behaviour using smartphone sensor data.
Previously, he acted as Chief Scientist and Vice President of Sentiance since joining in June 2014, during which he was responsible for building out the machine learning team at Sentiance, applying state-of-the-art academic research to real-life problems.
Vincent holds a PhD in machine learning and was awarded the MIT innovators under 35 award in 2017. He founded several startups in his past and has years of experience in both the technology industry and the world of academic research.
Being the driving force behind the Ethical AI task force within Sentiance, he is deeply involved in the process of providing tooling and education to ensure algorithmic fairness across the Sentiance platform.