Explainable AI: What, Why and How?

I will introduce the concept of Explainable AI, as conceptualised for instance by the DARPA XAI programme (as part of which we will be starting a project shortly), and talk about: (a) why this is being considered an important problem, (b) what questions are considered more pressing and (c) what sorts of tools might be brought to bear on this problem area. My talk will be informal in nature and will be biased towards my own perspective within this area. Lastly, my intention is to allow plenty of time for discussion - I am as interested in learning from others’ views on this topic as I am in sharing my own with the audience.