Augmenting Visual Question Answering with Semantic Frame Information in a Multitask Learning Approach | IEEE Conference Publication | IEEE Xplore