Generating Informative Dialogue Responses with Keywords-Guided Networks

Recently, open-domain dialogue systems have attracted growing attention. Most of them use the sequence-to-sequence (Seq2Seq) architecture to generate responses. However, traditional Seq2Seq-based models tend generic and safe responses, which are less informative, unlike human In this paper, we propose a simple but effective Keywords-guided Sequence-to-sequence model (KW-Seq2Seq) uses keywords information as guidance Specifically, given context, KW-Seq2Seq first decoder predict sequence topic keywords, then generates final response under them. Extensive experiments demonstrate that can facilitate produce more coherent, fluent yielding substantive gain in both automatic evaluation metrics.