The Dialog State Tracking Challenge

In a spoken dialog system, dialog state tracking deduces information about the user's goal as the dialog progresses, synthesizing evidence such as dialog acts over multiple turns with external data sources. Recent approaches have been shown to overcome ASR and SLU errors in some applications. However, there are currently no common testbeds or evaluation measures for this task, hampering progress. The dialog state tracking challenge seeks to address this by providing a heterogeneous corpus of 15K human-computer dialogs in a standard format, along with a suite of 11 evaluation metrics. The challenge received a total of 27 entries from 9 research groups. The results show that the suite of performance metrics cluster into 4 natural groups. Moreover, the dialog systems that benefit most from dialog state tracking are those with less discriminative speech recognition confidence scores. Finally, generalization is a key problem: in 2 of the 4 test sets, fewer than half of the entries out-performed simple baselines.

dstc2013.pdf
PDF file

In  Proceedings of the SIGDIAL 2013 Conference, Metz, France

Publisher  Association for Computational Linguistics

Details

TypeInproceedings
> Publications > The Dialog State Tracking Challenge