@InProceedings{zhang-EtAl:2018:C18-14,
  author    = {Zhang, Zhuosheng  and  Li, Jiangtong  and  Zhu, Pengfei  and  Zhao, Hai  and  Liu, Gongshen},
  title     = {Modeling Multi-turn Conversation with Deep Utterance Aggregation},
  booktitle = {Proceedings of the 27th International Conference on Computational Linguistics},
  month     = {August},
  year      = {2018},
  address   = {Santa Fe, New Mexico, USA},
  publisher = {Association for Computational Linguistics},
  pages     = {3740--3752},
  abstract  = {Multi-turn conversation understanding is a major challenge for building intelligent dialogue systems. This work focuses on retrieval-based response matching for multi-turn conversation whose related work simply concatenates the conversation utterances, ignoring the interactions among previous utterances for context modeling. In this paper, we formulate previous utterances into context using a proposed deep utterance aggregation model to form a fine-grained context representation. In detail, a self-matching attention is first introduced to route the vital information in each utterance. Then the model matches a response with each refined utterance and the final matching score is obtained after attentive turns aggregation. Experimental results show our model outperforms the state-of-the-art methods on three multi-turn conversation benchmarks, including a newly introduced e-commerce dialogue corpus.},
  url       = {http://www.aclweb.org/anthology/C18-1317}
}