@InProceedings{lison-bibauw:2017:W17-55,
  author    = {Lison, Pierre  and  Bibauw, Serge},
  title     = {Not All Dialogues are Created Equal: Instance Weighting for Neural Conversational Models},
  booktitle = {Proceedings of the 18th Annual SIGdial Meeting on Discourse and Dialogue},
  month     = {August},
  year      = {2017},
  address   = {Saarbrücken, Germany},
  publisher = {Association for Computational Linguistics},
  pages     = {384--394},
  abstract  = {Neural conversational models require substantial amounts of dialogue data to
	estimate their parameters and are therefore usually learned on large corpora
	such as chat forums or movie subtitles. These corpora are, however, often
	challenging to work with, notably due to their frequent lack of turn
	segmentation and the presence of multiple references external to the dialogue
	itself. This paper shows that these challenges can be mitigated by adding a
	weighting model into the architecture. The weighting model, which is itself
	estimated from dialogue data, associates each training example to a numerical
	weight that reflects its intrinsic quality for dialogue modelling. At training
	time, these sample weights are included into the empirical loss to be
	minimised. Evaluation results on retrieval-based models trained on movie and TV
	subtitles demonstrate that the inclusion of such a weighting model improves the
	model performance on unsupervised metrics.},
  url       = {http://aclweb.org/anthology/W17-5546}
}