PhD Student and Rotman Institute Member
I am first-year PhD student in philosophy and a new lab member excited to explore the relationship between machine and biological learning. I am working on a summer research project investigating how ecological psychology can help conceptualize the behaviour of machine learning algorithms, including their success at representation learning and their curious response to “adversarial” inputs. Do the uninterpretable features that deep artificial neural networks learn emerge as possibilities for action given a certain data geometry and a task? If so, J. J. Gibson’s notion of the “affordances” that constitute the interface between an animal and its environment may be apt. Machine “affordances” may help explain the success of these algorithms in a rich but restricted data environment.
Champion, H. (2021). "Representation without constraint: a critique of Poldrack’s account of object detection using artificial neural networks," Rotman Graduate Student Conference, Western University, May 2021.