Zhang Wen Yuxin Peng 0001 Multi-Level Knowledge Injecting for Visual Commonsense Reasoning. 1042-1054 2021 31 IEEE Trans. Circuits Syst. Video Technol. 3 https://doi.org/10.1109/TCSVT.2020.2991866 db/journals/tcsv/tcsv31.html#WenP21