Causal inference: some examples
Asohan Amarasingham, City College CUNY
The inference of causes from a systematic combination of observations and actions is central to the process of science. Could a framework for causal inference help? Practice may suggest not, given the evident fact that principles of causal inference do not routinely appear as a codified part of scientific education. To examine this question, I will work through several subtle but elementary examples in which careful thinking about causal definitions is illuminating and significantly affects causal conclusions. Some of the most interesting phenomena are almost immediately visible in toy examples, merely by setting up the problem the right way. I will discuss why we might expect such concepts to play an important role in day-to-day applications of machine learning. Time permitting, I will also reflect on how they have influenced my thinking about certain problems in neuroscience.