2

This model disentangles movement that can be quantified by keypoints (e.g., limb position) from subtler feature variations like orofacial movements. We introduce a novel VAE whose latent space decomposes into two orthogonal subspaces – one unsupervised subspace and one supervised subspace linearly predictive of labels (keypoints). The latent space additionally includes a context variable that predicts the video/subject identity.

Psychophysical presentation of ambiguous objects in context. Hierarchical Bayesian GLM and meta-analysis of subjects’ judgments.

A series of visual-serach experiments asking people to search for multiple features in parallel while ignoring distractors. The pattern of people’s attentional capture allowed us to characterize the cost of multi-tasking.