Virtual fMRI brown bag : February 4, 2022

Please join us for a talk given by Kate Storrs, the Alexander von Humboldt Research Fellow at Justus Liebig University in Giessen, Germany.

Supervised and Unsupervised Models of Vision

Abstract: Computational visual neuroscience has come a long way in the past 10 years. For the first time, we have fully explicit, image-computable models that can recognise objects with near-human accuracy, and predict brain activity in high-level visual regions. Diverse deep neural network architectures all predict ventral stream representations well, especially after reweighting model features to better match brain data. However, vision is far from explained. Our most successful models have been supervised to recognise objects in images using ground-truth labels for millions of examples. Brains have no such access to the ground truth, and must instead learn directly from sensory data. Unsupervised deep learning, in which networks learn statistical regularities in their data by compressing, extrapolating or predicting images and videos, presents a more ecologically feasible alternative. I will show evidence that unsupervised networks trained on environments of 3D rendered objects with varying shape, material and illumination, spontaneously come to encode these properties of the environment in their internal representations. More strikingly, they can predict, on an image-by-image basis, patterns of errors made by human observers. Unsupervised deep learning may provide a powerful framework for understanding how our perceptual dimensions arise.