Date : June 2016
Amin Sleimi and Claire Gardent. Generating Paraphrases from DBPedia using Deep Learning. WebNLG 2016, The Second International Workshop on NLG and the Semantic Web, September 6 2016, Edinburgh (Scotland).
Abstract : Recent deep learning approaches to Natural Language Generation mostly rely on sequence-to-sequence models. In these approaches, the input is treated as a sequence whereas in most cases, input to generation is either a tree or a graph. In this paper, we describe an experiment which shows that enriching a sequential input with structural information (i) improves results and (ii) helps support the generation of paraphrases.