Xiangyu Li's paper "Natural Answer Generation via Graph Transformer" has been accepted by APWeb-WAIM 2020.

　　Natural Answer Generation (NAG), which generates natural answer sentences for the given question, has received much attention in recent years. Compared with traditional QA systems, NAG could offer specific entities fluently and naturally, which is more user-friendly in the real world. However, existing NAG systems usually utilize simple retrieval and embedding mechanism, which is hard to tackle complex questions. They suffer issues containing knowledge insufficiency, entity ambiguity, and especially poor expressiveness during generation. To address these challenges, we propose an improved knowledge extractor to retrieve supporting graphs from the knowledge base, and an extending graph transformer to encode the supporting graph, which considers global and variable information as well as the communication path between entities. By testing the data set in the film field, the model in this paper can provide natural language answers to complex questions, and the evaluation index is higher than previous work.