Shades of BLEU, Flavours of Success: The Case of MultiWOZ

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F21%3A10440567" target="_blank" >RIV/00216208:11320/21:10440567 - isvavai.cz</a>

Result on the web

<a href="https://aclanthology.org/2021.gem-1.4/" target="_blank" >https://aclanthology.org/2021.gem-1.4/</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.18653/v1/2021.gem-1.4" target="_blank" >10.18653/v1/2021.gem-1.4</a>

Result language

angličtina

Original language name

Shades of BLEU, Flavours of Success: The Case of MultiWOZ

Original language description

The MultiWOZ dataset (Budzianowski et al.,2018) is frequently used for benchmarking context-to-response abilities of task-oriented dialogue systems. In this work, we identify inconsistencies in data preprocessing and reporting of three corpus-based metrics used on this dataset, i.e., BLEU score and Inform &amp; Success rates. We point out a few problems of the MultiWOZ benchmark such as unsatisfactory preprocessing, insufficient or under-specified evaluation metrics, or rigid database. We re-evaluate 7 end-to-end and 6 policy optimization models in as-fair-as-possible setups, and we show that their reported scores cannot be directly compared. To facilitate comparison of future systems, we release our stand-alone standardized evaluation scripts. We also give basic recommendations for corpus-based benchmarking in future works.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

S - Specificky vyzkum na vysokych skolach<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Publication year

2021

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Proceedings of the 1st Workshop on Natural Language Generation, Evaluation, and Metrics (GEM 2021)

ISBN

978-1-954085-67-1

ISSN

—

e-ISSN

—

Number of pages

13

Pages from-to

34-46

Publisher name

Association for Computational Linguistics

Place of publication

Stroudsburg, PA, USA

Event location

Online

Event date

Aug 1, 2021

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—