Shang-Yu Su, Kai-Ling Lo, Yi-Ting Yeh, and Yun-Nung Chen.
Full Paper: arXiv
Natural language generation (NLG) is a critical component in spoken dialogue systems.
Classic NLG can be divided into two phases:
- sentence planning: deciding on the overallsentence structure
- surface realization: determining specific word forms and flatteningthe sentence structure into a string.
Many simple NLG models are based on recurrent neural networks (RNN) and sequence-to-sequence(seq2seq) model, which basically contains an encoder-decoder structure; these NLG models generate sentences from scratch by jointly optimizing sentence planning and surface realization using a simple cross entropy loss training criterion.
However, the simple encoder-decoder architecture usually suffers from generating complex and long sentences, becausethe decoder has to learn all grammar and diction knowledge.
This paper introduces a hierarchical decoding NLG model based on linguistic patterns in different levels, and shows that the proposed method outperforms the traditional one with a smaller model size. Furthermore, the design of the hierarchical decoding is flexible and easily-extensible in various NLG systems