Natural Answer Generation with QA Pairs Using Sequence to Sequence Model

摘要

Generating natural answer is an important task in question answering systems. QA systems are usually designed to response right answers as well as friendly natural sentences to users. In this paper, we apply Sequence to Sequence Model based on LSTM in Chinese natural answer generation. By taking advantage of LSTM in context inference, we build such a model that can learn from question-answer pairs (One question sentence and one answer entity pair), and finally generate natural answer sentences. Experiments show that our model has achieved good effectiveness.