Memory State Tracker: A Memory Network based Dialogue State Tracker

被引：0

作者：

Wang, Di ^{[1
]}

O'Keefe, Simon ^{[1
]}

机构：

[1] Univ York, Dept Comp Sci, York, N Yorkshire, England

来源：

ICAART: PROCEEDINGS OF THE 13TH INTERNATIONAL CONFERENCE ON AGENTS AND ARTIFICIAL INTELLIGENCE - VOL 1 | 2021年

关键词：

Dialogue State Tracker; Memory Network;

D O I：

10.5220/0010385705330538

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Dialogue State Tracking (DST) is a core component towards task oriented dialogue system. It fills manually-set slots at each turn of an utterance, which indicate the current topics or user requirement. In this work we propose a memory based state tracker that includes a memory encoder which encodes the dialogue history into a memory vector, and then connects to a pointer network which makes predictions. Our model reached a joint goal accuracy of 49.16% on MultiWOZ 2.0 data set (Budzianowski et al., 2018) and 47.27% on MultiWOZ 2.1 data set (Eric et al., 2019), outperforming the benchmark result.

引用

页码：533 / 538

页数：6

共 19 条

[11]

Pennington J., 2014, P 2014 C EMP METH NA, P1532, DOI DOI 10.3115/V1/D14-1162

[12]

Poupart P, 2019, ARXIV PREPRINT ARXIV

[13]

Qiu L, 2019, 57TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2019), P6140

[14]

Ramadan O., 2018, ARXIV PREPRINT ARXIV

[15]

Rosenfeld A, 2014, AAMAS'14: PROCEEDINGS OF THE 2014 INTERNATIONAL CONFERENCE ON AUTONOMOUS AGENTS & MULTIAGENT SYSTEMS, P525

[16]

Sukhbaatar S., 2015, P ADV NEUR INF PROC, P2440

[17]

Wu Chenfei, 2019, ARXIV PREPRINT ARXIV

[18]

Xu P., 2018, ARXIV PREPRINT ARXIV

[19]

Zhong V, 2018, PROCEEDINGS OF THE 56TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL), VOL 1, P1458

← 1 2 →