Incremental Joint Modelling for Dialogue State Tracking

Anh Duong Trinh; Robert J. Ross; John D. Kelleher

doi:10.21427/cktn-wm29

Incremental Joint Modelling for Dialogue State Tracking

Anh Duong Trinh, Robert J. Ross, John D. Kelleher

School of Computer Science

Research output: Contribution to conference › Paper › peer-review

Abstract

Dialogue State Tracking is an important task in dialogue management as it provides a mechanism to monitor dialogue contributions. In this paper we introduce an Incremental Joint Model as a new approach to the task. Our tracker is capable of incrementally tracking Dialogue States. We base our model and analysis on the datasets provided in the Second Dialogue State Tracking Challenge (DSTC2). Our early stage evaluations are based on comparisons of our tracker with both the baseline model provide by the DSTC2 and also LecTrack: a state-of-the-art incremental LSTM-based tracker. The main finding of our experiments is that moving from an utterance based to incremental word based tracker results in better performance for our RNN based joint task models.

Original language	English
DOIs	https://doi.org/10.21427/cktn-wm29
Publication status	Published - 2017
Event	SEMDIAL 2017 (SaarDial) Workshop on the Semantics and Pragmatics of Dialogue - Saarbrucken, Germany Duration: 15 Aug 2017 → 17 Aug 2017

Conference

Conference	SEMDIAL 2017 (SaarDial) Workshop on the Semantics and Pragmatics of Dialogue
Country/Territory	Germany
City	Saarbrucken
Period	15/08/17 → 17/08/17

Keywords

Dialogue State Tracking
dialogue management
Incremental Joint Model
Second Dialogue State Tracking Challenge
DSTC2
LecTrack
incremental word based tracker
RNN based joint task models

Access to Document

10.21427/cktn-wm29Licence: CC BY-NC-SA

Cite this

@conference{7e80c59defaa44d6bc865d904168e840,

title = "Incremental Joint Modelling for Dialogue State Tracking",

abstract = "Dialogue State Tracking is an important task in dialogue management as it provides a mechanism to monitor dialogue contributions. In this paper we introduce an Incremental Joint Model as a new approach to the task. Our tracker is capable of incrementally tracking Dialogue States. We base our model and analysis on the datasets provided in the Second Dialogue State Tracking Challenge (DSTC2). Our early stage evaluations are based on comparisons of our tracker with both the baseline model provide by the DSTC2 and also LecTrack: a state-of-the-art incremental LSTM-based tracker. The main finding of our experiments is that moving from an utterance based to incremental word based tracker results in better performance for our RNN based joint task models.",

keywords = "Dialogue State Tracking, dialogue management, Incremental Joint Model, Second Dialogue State Tracking Challenge, DSTC2, LecTrack, incremental word based tracker, RNN based joint task models",

author = "Trinh, \{Anh Duong\} and Ross, \{Robert J.\} and Kelleher, \{John D.\}",

year = "2017",

doi = "10.21427/cktn-wm29",

language = "English",

note = "SEMDIAL 2017 (SaarDial) Workshop on the Semantics and Pragmatics of Dialogue ; Conference date: 15-08-2017 Through 17-08-2017",

}

TY - CONF

T1 - Incremental Joint Modelling for Dialogue State Tracking

AU - Trinh, Anh Duong

AU - Ross, Robert J.

AU - Kelleher, John D.

PY - 2017

Y1 - 2017

N2 - Dialogue State Tracking is an important task in dialogue management as it provides a mechanism to monitor dialogue contributions. In this paper we introduce an Incremental Joint Model as a new approach to the task. Our tracker is capable of incrementally tracking Dialogue States. We base our model and analysis on the datasets provided in the Second Dialogue State Tracking Challenge (DSTC2). Our early stage evaluations are based on comparisons of our tracker with both the baseline model provide by the DSTC2 and also LecTrack: a state-of-the-art incremental LSTM-based tracker. The main finding of our experiments is that moving from an utterance based to incremental word based tracker results in better performance for our RNN based joint task models.

AB - Dialogue State Tracking is an important task in dialogue management as it provides a mechanism to monitor dialogue contributions. In this paper we introduce an Incremental Joint Model as a new approach to the task. Our tracker is capable of incrementally tracking Dialogue States. We base our model and analysis on the datasets provided in the Second Dialogue State Tracking Challenge (DSTC2). Our early stage evaluations are based on comparisons of our tracker with both the baseline model provide by the DSTC2 and also LecTrack: a state-of-the-art incremental LSTM-based tracker. The main finding of our experiments is that moving from an utterance based to incremental word based tracker results in better performance for our RNN based joint task models.

KW - Dialogue State Tracking

KW - dialogue management

KW - Incremental Joint Model

KW - Second Dialogue State Tracking Challenge

KW - DSTC2

KW - LecTrack

KW - incremental word based tracker

KW - RNN based joint task models

U2 - 10.21427/cktn-wm29

DO - 10.21427/cktn-wm29

M3 - Paper

T2 - SEMDIAL 2017 (SaarDial) Workshop on the Semantics and Pragmatics of Dialogue

Y2 - 15 August 2017 through 17 August 2017

ER -

Incremental Joint Modelling for Dialogue State Tracking

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this