Multi-domain Dialog State Tracking using Recurrent Neural Networks
release_oad6amofujeiffmqfjsqfcefsm
by
Nikola Mrkšić, Diarmuid Ó Séaghdha, Blaise Thomson, Milica
Gašić, Pei-Hao Su, David Vandyke, Tsung-Hsien Wen, Steve Young
2015
Abstract
Dialog state tracking is a key component of many modern dialog systems, most
of which are designed with a single, well-defined domain in mind. This paper
shows that dialog data drawn from different dialog domains can be used to train
a general belief tracking model which can operate across all of these domains,
exhibiting superior performance to each of the domain-specific models. We
propose a training procedure which uses out-of-domain data to initialise belief
tracking models for entirely new domains. This procedure leads to improvements
in belief tracking performance regardless of the amount of in-domain data
available for training the model.
In text/plain
format
Archived Files and Locations
application/pdf 166.1 kB
file_wpi3wnp3lzg6tkekhv37fzexby
|
arxiv.org (repository) web.archive.org (webarchive) |
1506.07190v1
access all versions, variants, and formats of this works (eg, pre-prints)