As AI is increasingly relied upon to make decisions that govern our lives, it is vital that we understand how to ensure that the decisions these algorithms are making are fair.

This talk, delivered by Faculty's Research & Development Lead Tom Begley, will discuss how we can determine whether a model is fair or not, and how we can try to correct unfair models.

Fairness is definitely not a "solved" problem and there is a lot of ongoing research. Throughout the talk Tom will point out problems of current fairness methods when applied to real-world problems and outline research efforts that are being conducted now.

The Faculty Research Lab, in collaboration with top universities, explores the frontier of AI through the publication of research papers and the development of novel technology. We consider our research efforts in terms of two categories: improving the safety of machine learning algorithms (AI Safety) and advancing their capabilities (ML Research).

This talk will cover:

• Why fairness is important (hopefully we don't have to convince you too much…)

• How to define fairness in the first place, and the trade-offs between different notions of fairness

• How fairness can be achieved

• What popular methods are out there at the moment? What are their advantages and drawbacks?

No AI/coding experience necessary for this talk.

Event Timings:

18:00 Event opens

18:30 Talk starts

19:30 Networking, beer and pizza

20:30 Event close