XAI-Explainable artificial intelligence

被引：961

作者：

Gunning, David ^{[1
,7
]}

Stefik, Mark ^{[2
]}

Choi, Jaesik ^{[3
]}

Miller, Timothy ^{[4
]}

Stumpf, Simone ^{[5
]}

Yang, Guang-Zhong ^{[6
]}

机构：

[1] DARPA, 675 North Randolph St, Arlington, VA 22201 USA

[2] Palo Alto Res Ctr, 3333 Coyote Hill Rd, Palo Alto, CA 94304 USA

[3] Korea Adv Inst Sci & Technol, Grad Sch Artificial Intelligence, 291 Daehak Ro, Daejeon 34141, South Korea

[4] Univ Melbourne, Sch Comp & Informat Syst, Melbourne, Vic 3010, Australia

[5] City Univ London, Sch Math Comp Sci & Engn, Ctr HCI Design, London EC1V 0HB, England

[6] Shanghai Jiao Tong Univ, Inst Med Robot, Shanghai, Peoples R China

[7] Facebook AI Res, 770 Broadway, New York, NY 10003 USA

来源：

SCIENCE ROBOTICS | 2019年 / 4卷 / 37期

关键词：

D O I：

10.1126/scirobotics.aay7120

中图分类号：

TP24 [机器人技术];

学科分类号：

080202 ; 1405 ;

摘要：

Recent successes in machine learning (ML) have led to a new wave of artificial intelligence (AI) applications that offer extensive benefits to a diverse range of fields. However, many of these systems are not able to explain their autonomous decisions and actions to human users. Explanations may not be essential for certain AI applications, and some AI researchers argue that the emphasis on explanation is misplaced, too difficult to achieve, and perhaps unnecessary. However, for many critical applications in defense, medicine, finance, and law, explanations are essential for users to understand, trust, and effectively manage these new, artificially intelligent partners [see recent reviews (1-3)]. Copyright © 2019 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works

引用

页数：2

共 9 条

[1]

[Anonymous], 1991, PERSPECTIVES SOCIALL, DOI DOI 10.1037/10096-006

[2] Intelligibility and accountability: Human considerations in context-aware systems [J].

Bellotti, V ;

Edwards, K .

HUMAN-COMPUTER INTERACTION, 2001, 16 (2-4) :193-212

[3]

Biran O., 2017, IJCAI 17 WORKSHOP EX, P8

[4]

Escalante H.J., 2018, EXPLAINABLE INTERPRE, Vfirst

[5]

Gunning D., EXPLAINABLE ARTIFICI

[6]

Kulesza T., 2015, P 20 INT C INTELLIGE, P126, DOI DOI 10.1145/2678025.2701399

[7] Explanation in artificial intelligence: Insights from the social sciences [J].

Miller, Tim .

ARTIFICIAL INTELLIGENCE, 2019, 267 :1-38

[8]

Samek W., 2019, Springer Nature, V11700

[9]

Wang D., 2019, P 2019 CHI C HUM FAC, P1

← 1 →