SAFETYKIT: First Aid for Measuring Safety in Open-domain Conversational Systems

被引：0

作者：

Dinan, Emily ^{[1
]}

Abercrombie, Gavin ^{[2
]}

Bergman, A. Stevie ^{[3
]}

Spruit, Shannon ^{[4
]}

Hovy, Dirk ^{[5
]}

Boureau, Y-Lan ^{[1
]}

Rieser, Verena ^{[2
,6
]}

机构：

[1] Facebook AI Res, Menlo Pk, CA 94025 USA

[2] Heriot Watt Univ, Edinburgh, Midlothian, Scotland

[3] Facebook, Responsible AI, Menlo Pk, CA USA

[4] Populytics, Leiden, Netherlands

[5] Bocconi Univ, Milan, Italy

[6] Alana AI, London, England

来源：

PROCEEDINGS OF THE 60TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2022), VOL 1: (LONG PAPERS) | 2022年

基金：

英国工程与自然科学研究理事会; 欧洲研究理事会;

关键词：

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

The social impact of natural language processing and its applications has received increasing attention. In this position paper, we focus on the problem of safety for end-to-end conversational AI. We survey the problem landscape therein, introducing a taxonomy of three observed phenomena: the INSTIGATOR, YEASAYER, and IMPOSTOR effects. We then empirically assess the extent to which current tools can measure these effects and current systems display them. We release these tools as part of a "first aid kit" (SAFETYKIT) to quickly assess apparent safety concerns. Our results show that, while current tools are able to provide an estimate of the relative safety of systems in various settings, they still have several shortcomings. We suggest several future directions and discuss ethical considerations.

引用

下载

页码：4113 / 4133

页数：21

共 50 条

[1] Offline and Online Satisfaction Prediction in Open-Domain Conversational Systems
Choi, Jason Ingyu
Ahmadvand, Ali
Agichtein, Eugene
PROCEEDINGS OF THE 28TH ACM INTERNATIONAL CONFERENCE ON INFORMATION & KNOWLEDGE MANAGEMENT (CIKM '19), 2019, : 1281 - 1290
[2] Learning to Ask Questions in Open-domain Conversational Systems with Typed Decoders
Wang, Yansen
Liu, Chenyi
Huang, Minlie
Nie, Liqiang
PROCEEDINGS OF THE 56TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL), VOL 1, 2018, : 2193 - 2203
[3] MERCY: Multiple Response Ranking Concurrently in Realistic Open-Domain Conversational Systems
Ghazarian, Sarik
Hedayatnia, Behnam
Jin, Di
Liu, Sijia
Peng, Violet
Liu, Yang
Hakkani-Tur, Dilek
24TH MEETING OF THE SPECIAL INTEREST GROUP ON DISCOURSE AND DIALOGUE, SIGDIAL 2023, 2023, : 615 - 631
[4] DISTRIBUTED OPEN-DOMAIN CONVERSATIONAL UNDERSTANDING FRAMEWORK WITH DOMAIN INDEPENDENT EXTRACTORS
Li, Qi
Tur, Gokhan
Hakkani-Tur, Dilek
Li, Xiang
Paek, Tim
Gunawardana, Asela
Quirk, Chris
2014 IEEE WORKSHOP ON SPOKEN LANGUAGE TECHNOLOGY SLT 2014, 2014, : 566 - 571
[5] Order-Sensitive Keywords Based Response Generation in Open-Domain Conversational Systems
Zhu, Qingfu
Zhang, Weinan
Cui, Lei
Liu, Ting
ACM TRANSACTIONS ON ASIAN AND LOW-RESOURCE LANGUAGE INFORMATION PROCESSING, 2020, 19 (02)
[6] Knowledge Graph Enabled Open-Domain Conversational Question Answering
Oduro-Afriyie, Joel
Jamil, Hasan
FLEXIBLE QUERY ANSWERING SYSTEMS, FQAS 2023, 2023, 14113 : 63 - 76
[7] State-of-the-Art in Open-Domain Conversational AI: A Survey
Adewumi, Tosin
Liwicki, Foteini
Liwicki, Marcus
INFORMATION, 2022, 13 (06)
[8] Contextual Dialogue Act Classification for Open-Domain Conversational Agents
Ahmadvand, Ali
Choi, Jason Ingyu
Agichtein, Eugene
PROCEEDINGS OF THE 42ND INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL (SIGIR '19), 2019, : 1273 - 1276
[9] TopiOCQA: Open-domain Conversational Question Answering with Topic Switching
Adlakha, Vaibhav
Dhuliawala, Shehzaad
Suleman, Kaheer
de Vries, Harm
Reddy, Siva
TRANSACTIONS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, 2022, 10 : 468 - 483
[10] Dynamic Graph Reasoning for Conversational Open-Domain Question Answering
Li, Yongqi
Li, Wenjie
Nie, Liqiang
ACM TRANSACTIONS ON INFORMATION SYSTEMS, 2022, 40 (04)

← 1 2 3 4 5 →