Structured Discriminative Model For Dialog State Tracking

Many dialog state tracking algorithms have been limited to generative modeling due to the influence of the Partially Observable Markov Decision Process framework. Recent analyses, however, raised fundamental questions on the effectiveness of the generative formulation. In this paper, we present a structured discriminative model for dialog state tracking as an alternative. Unlike generative models, the proposed method affords the incorporation of features without having to consider dependencies between observations. It also provides a flexible mechanism for imposing relational constraints. To verify the effectiveness of the proposed method, we applied it to the Let’s Go domain (Raux et al., 2005). The results show that the proposed model is superior to the baseline and generative model-based systems in accuracy, discrimination, and robustness to mismatches between training and test datasets.

[1]  Milica Gasic,et al.  The Hidden Information State model: A practical framework for POMDP-based spoken dialogue management , 2010, Comput. Speech Lang..

[2]  Alexander I. Rudnicky,et al.  A “K Hypotheses + Other” Belief Updating Model , 2006 .

[3]  Maxine Eskénazi,et al.  Spoken Dialog Challenge 2010: Comparison of Live and Control Test Results , 2011, SIGDIAL Conference.

[4]  Jason D. Williams.  A critical analysis of two statistical spoken dialog systems in public use , 2012, 2012 IEEE Spoken Language Technology Workshop (SLT).

[5]  Amir Globerson,et al.  An LP View of the M-best MAP problem , 2009, NIPS.

[6]  Maxine Eskénazi,et al.  Let's go public! taking a spoken dialog system to the real world , 2005, INTERSPEECH.

[7]  I JordanMichael,et al.  Graphical Models, Exponential Families, and Variational Inference , 2008 .

[8]  Milica Gasic,et al.  Effective handling of dialogue state in the hidden information state POMDP-based dialogue manager , 2011, TSLP.

[9]  Ben Taskar,et al.  An Introduction to Conditional Random Fields for Relational Learning , 2007 .

[10]  Maxine Eskénazi,et al.  POMDP-based Let's Go system for spoken dialog challenge , 2012, 2012 IEEE Spoken Language Technology Workshop (SLT).

[11]  Jianfeng Gao,et al.  Scalable training of L1-regularized log-linear models , 2007, ICML '07.

[12]  Yi Ma,et al.  Efficient Probabilistic Tracking of User Goal and Dialog History for Spoken Dialog Systems , 2011, INTERSPEECH.

[13]  Ben Taskar,et al.  Introduction to statistical relational learning , 2007 .

[14]  Jason Williams.  An Empirical Evaluation of a Statistical Dialog System in Public Use , 2011, SIGDIAL Conference.

[15]  Steve J. Young,et al.  Bayesian update of dialogue state: A POMDP framework for spoken dialogue systems , 2010, Comput. Speech Lang..

[16]  Michael I. Jordan,et al.  Graphical Models, Exponential Families, and Variational Inference , 2008, Found. Trends Mach. Learn..

[17]  Y. Weiss,et al.  Finding the M Most Probable Configurations using Loopy Belief Propagation , 2003, NIPS 2003.

[18]  Steve J. Young,et al.  Partially observable Markov decision processes for spoken dialog systems , 2007, Comput. Speech Lang..

[19]  David J. Spiegelhalter,et al.  Local computations with probabilities on graphical structures and their application to expert systems , 1990 .

[20]  Jason D. Williams.  Incremental partition recombination for efficient tracking of multiple dialog states , 2010, 2010 IEEE International Conference on Acoustics, Speech and Signal Processing.

[21]  Maxine Eskénazi,et al.  Exploiting Machine-Transcribed Dialog Corpus to Improve Multiple Dialog States Tracking Methods , 2012, SIGDIAL Conference.

[22]  Milica Gasic,et al.  Bayesian dialogue system for the Let's Go Spoken Dialogue Challenge , 2010, 2010 IEEE Spoken Language Technology Workshop.

[23]  Milica Gasic,et al.  Parameter learning for POMDP spoken dialogue models , 2010, 2010 IEEE Spoken Language Technology Workshop.

[24]  Maxine Eskénazi,et al.  Recipe For Building Robust Spoken Dialog State Trackers: Dialog State Tracking Challenge System Description , 2013, SIGDIAL Conference.