Machine learning and the acceleration of discovery

ePODstemology is about popularising the genuinely new ways of thinking emerging from the pathbreaking research of young scholars. There are few fields that represent this agenda more than machine learning, a branch of computer science and statistics that promises to dramatically accelerate the pace of scientific discovery, crack open hard questions that have bedeviled humanity for decades, and even crack open our minds with whole new ways of understanding our world. In this episode, regular host Dr Mark Fabian from the universities of Tasmania and Cambridge is joined by Dr David Watson, a postdoc at University College London making voluminous contributions to the machine learning literature. Their discussion ranges over what machine learning, why deep neural networks are so hot right now, how they are being applied in areas as diverse as protein folding and x-ray imaging, and their potential to transform society, for both good and ill. https://dswatson.github.io/The rhetoric and reality of anthropomorphism in artificial intelligence. https://link.springer.com/article/10.1007/s11023-019-09506-6Local explanations via necessity and sufficiency: Unifying theory and practice. https://link.springer.com/article/10.1007/s11023-022-09598-7 Mass Effect Full Game Movie: https://www.youtube.com/watch?v=B5dSRaeUcFc Joy, Bladerunner 2049: https://www.youtube.com/watch?v=RL0gX1_NWTkKazuo Ishiguro, Klara and the SunIan McEwan, Machines Like MeAthey and Imbens – Machine learning methods economists should know about. https://www.annualreviews.org/doi/abs/10.1146/annurev-economics-080217-053433