You Don’t Need Strong Assumptions: Visual Representation Learning via Temporal Differences | Ninad Daithankar et al. | ResearchPod