Common Ground Day: Ethics and Machine Learning

Common Ground Day: Ethics and Machine Learning

Claire Benn, Dan Steinberg, and Lachlan McCalman

This day-long course at the Coral Bell School brought together research fellows from HMI and the Gradient Institute to exchange knowledge on both ethics and machine learning. Claire Benn presented the basics of ethics: different normative domains; the main normative theories such as consequentialism, deontology, virtue ethics; some important normative dimensions, such as the difference between good, better, best, ought, required, permitted, etc; and introduced some basics concerning incommensurability and moral dilemmas.

Dan and Lachy from the Gradient Institute introduced the core elements of machine learning, including the different kinds of tasks (supervised, unsupervised and reinforcement), coordinates, functions, conditional probability, and optimisation. They demonstrated the differences between regression and classification and explored some choices concerning fit, validation, loss functions and different kinds of errors. They concluded with a discussion of causal modelling and modelling uncertainty.