1 paper
Anthony Colas, Seokhwan Kim, Franck Dernoncourt +3
Despite the number of currently available datasets on video question answering, there still remains a need for a dataset involving multi-step and non-factoid answers. Moreover, rel…