Acquiring linguistic argument structure from multimodal input using attentive focus | IEEE Conference Publication | IEEE Xplore