Modern Machine Learning, Hysteria, Hope and Hype

In this lecture, Dr Thompson will explain why it is that Machine Learning has been able to create models of human language that are good enough to talk back to us, and why it’s been able to create models of proteins that can predict previously unknown structures, and of course – all sorts of other models. It’s also important to understand how these kinds of models can be used, because understanding where they come from, how they came to be, and how they can be manipulated shines a light on what they can really do, and how Machine Learning must further develop if it’s going to really allow us to make humans happier, healthier and more harmless than is the current case. Even more significantly, an intuitive understanding of the principals of modern machine learning is useful in parsing hype from reality, and also seeing where the technology of modern AI can be misused out of malice or mischief.