Translating responsible AI principles to create VerifyML. User feedback, design decisions and architecture choices in creating our responsible AI solution.
Fairness is messy and complicated. Attempts to distil it down to a single metric is unhelpful and counter-productive. As business owners and model developers we should embrace the struggle in trying to apply fairness in artificial intelligence and data analytics models.
An explanation of the challenges of graph anonymisation and the difficulty of striking a balance between usefulness and anonymity. Written as a response to Singapore's TraceTogether privacy saga