论文部分内容阅读
In this paper, we develop a novel global-attentionbased neural network(GANN) for vision language intelligence,specifically, image captioning(language description of a given image). As many previous works, the encoder-decoder framework is adopted in our