Questions for "CLEVRER: Collision Events for Video Representation and Reasoning" ----------------------------------------------------------------------------------------------- Please send your answers to: schroepp@cs.uni-freiburg.de by 10:15 on 02.07.2020 1. What is the problem with previous video reasoning benchmarks and what is different regarding the focus of the CLEVRER dataset? (2-3 sentences, explain in your own words) 2. What data does CLEVRER contain, what is the task of the benchmark and which metrics are used for evaluation? (3 sentences) 3. Can the proposed NS-DR model be applied to other video reasoning benchmarks that potentially have differently structured questions? (2-3 sentences)