Skip to main navigation Skip to search Skip to main content

An auto-encoder for learning conversation representation using LSTM

  • Harbin Institute of Technology Shenzhen

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this paper, an auto-encoder is proposed to learn conversation representation. First, the long short term memory (LSTM) neural network is used to encode the sequence of sentences in a conversation. The interactive context is encoded into a fixed-length vector. Then, through the LSTM-decoder, the learnt representation is used to reconstruct the sentence vectors of a conversation. To train our model, we construct one corpus with 32,881 conversations from the online shopping platform. Finally, experiments on topic recognition task demonstrate the effectiveness of the proposed auto-encoder on learning conversation representation, especially when training data of topic recognition is relatively small.

Original languageEnglish
Title of host publicationNeural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings
EditorsWeng Kin Lai, Qingshan Liu, Tingwen Huang, Sabri Arik
PublisherSpringer Verlag
Pages310-317
Number of pages8
ISBN (Print)9783319265315
DOIs
StatePublished - 2015
Externally publishedYes
Event22nd International Conference on Neural Information Processing, ICONIP 2015 - Istanbul, Turkey
Duration: 9 Nov 201512 Nov 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9489
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on Neural Information Processing, ICONIP 2015
Country/TerritoryTurkey
CityIstanbul
Period9/11/1512/11/15

Keywords

  • Auto-encoder
  • Conversation representation
  • LSTM

Fingerprint

Dive into the research topics of 'An auto-encoder for learning conversation representation using LSTM'. Together they form a unique fingerprint.

Cite this