In this paper, we propose a novel pretraining-based encoder-decoder framework, which can generate the output sequence based on the input sequence in a two-stage manner. For the encoder of our model, we encode the input sequence into context representations using BERT. For the decoder, there are two stages in our model, in the first stage, we use a Transformer-based decoder to generate a draft output sequence. In the second stage, we mask each word of the draft sequence and feed it to BERT, then by combining the input sequence and the draft representation generated by BERT, we use a Transformer-based decoder to predict the refined word for each masked position. To the best of our knowledge, our approach is the first method which applies the BERT into text generation tasks. As the first step in this direction, we evaluate our proposed method on the text summarization task. Experimental results show that our model achieves new state-of-the-art on both CNN/Daily Mail and New York Times datasets.

In this work, we focus on complex question semantic parsing and propose a novel Hierarchical Semantic Parsing (HSP) method, which utilizes the decompositionality of complex questions for semantic parsing. Our model is designed within a three-stage parsing architecture based on the idea of decomposition-integration. In the first stage, we propose a question decomposer which decomposes a complex question into a sequence of sub-questions. In the second stage, we design an information extractor to derive the type and predicate information of these questions. In the last stage, we integrate the generated information from previous stages and generate a logical form for the complex question. We conduct experiments on COMPLEXWEBQUESTIONS which is a large scale complex question semantic parsing dataset, results show that our model achieves significant improvement compared to state-of-the-art methods.