2018/03/29 by Unnat Jain, Jain, Unnat, Svetlana Lazebnik +3
Computer Science · #Advanced Image and Video Retrieval Techniques #Computation and Language (cs.CL) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Multimodal Machine Learning Applications #Video Analysis and Summarization
paper · pdf · doi:10.48550/arxiv.1803.11186
openalex publication_date 2018/03/29 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28
Human conversation is a complex mechanism with subtle nuances. It is hence an\nambitious goal to develop artificial intelligence agents that can participate\nfluently in a conversation. While we are still far from achieving this goal,\nrecent progress in visual question answering, image captioning, and visual\nquestion generation shows that dialog systems may be realizable in the not too\ndistant future. To this end, a novel dataset was introduced recently and\nencouraging results were demonstrated, particularly for question answering. In\nthis paper, we demonstrate a simple symmetric discriminative baseline, that can\nbe applied to both predicting an answer as well as predicting a question. We\nshow that this method performs on par with the state of the art, even memory\nnet based methods. In addition, for the first time on the visual dialog\ndataset, we assess the performance of a system asking questions, and\ndemonstrate how visual dialog can be generated from discriminative question\ngeneration and question answering.\n