广告

Explainable AI for Entertainment: Issues on Video on Demand Platforms

会议论文
  • 554Downloads
的一部分Lecture Notes in Networks and Systems书系列(LNN,第223卷)

Abstract

With the proliferation of Artificial Intelligence-based systems, several questions arise involving ethical principles. In addition, the human-centered approach takes the focus on the user experience with these systems and studies user needs. A growing issue is the relationship between the transparency of these systems and the trust of users, since most systems are considered black-boxes. In this scenario, the Explainable Artificial Intelligence (XAI) emerges, with the proposal to explain the rationale of the decision making of the algorithms. XAI then starts to gain space in systems that involve high risk, such as health. Our research aims to discuss the importance of transparency to improve the user experience with recommendation mechanisms for entertainment, such as Video on Demand (VoD) platforms. In addition, we intent to raise the adjacent consequences of including XAI, such as improving the control and trust of VoD platforms. For this, we conducted an exploratory research method named Directed Storytelling. The study was conducted with thirty-one participants, all users of VoD platforms, regardless of time and frequency of use of this kind of systems. We note that people understand that there is an automated mechanism making recommendations for content in a personalized way for them, based on their browsing history, but the rules are not explicit. Thus, many users are suspicious of being manipulated by the system’s recommendations and resort to external recommendations, such as tips from third parties or Internet searches through specialized channels.

关键词

可解释的ai(xai) 人工智能(AI) 用户体验(UX) 机器学习(ml) 视频点播(VOD)平台

笔记

Acknowledgments

这项研究的部分经费由Coordenacaode Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001.

References

  1. 1。
    Holzinger,a .:从机器学习解释为可解释的ai。在:DISA 2018 - IEEE智能世界研讨会,用于系统和机器课程,PP。55-66(2018)谷歌学术
  2. 2。
    Oxborought, C., Cameron, E., Rao, A., Wetermann, C.: Explainable AI: driving business value throught greater understanding. PwC-UK (2018)谷歌学术
  3. 3。
    Ribera,M.,Lapedriza,a .:我们可以做更好的解释吗?以用户为中心的解释AI的提议。在:ACM IUI 2019讲习班,洛杉矶,美国的联合诉讼程序,3月20日,P。7. ACM,纽约(2019)谷歌学术
  4. 4。
    Samek, W., Wiegand, T., Müller, K.-R.: Explainable artificial intelligence: understanding, visualizing and interpreting deep learning models (2017)谷歌学术
  5. 5。
    Gade, K., Geyik, S., Kenthapadi, K., et al.: Explainable AI in industry: practical challenges and lessons learned. In: Companion Proceedings of the Web Conference 2020, pp. 303–304. ACM, New York (2020)谷歌学术
  6. 6。
    Samih,A.,Adadi,A.,Berrada,M。:朝着基于知识的可解释的推荐系统。:ACM诉讼国际会议(2019年)谷歌学术
  7. 7。
    徐,F.,Uszkoreit,H.,Du,Y.等:解释了AI:关于历史,研究领域,方法和挑战的简要调查。在:唐,J.,Kan,M.-Y.,Zhao,D。等。(EDS。)自然语言处理和中式计算:第8届CCF国际会议,NLPCC 2019年10月9日至14日,2019年10月议程,第二部分,第563-574部分。beplay登入Springer,Dunhuang(2019)谷歌学术
  8. 8。
    Aggarwal, C.C.: Recommender Systems: The Textbook. Springer, Yorktown Heights (2016)十字架谷歌学术
  9. 9.
    Konstan, J.A., Riedl, J.: Recommender systems: from algorithms to user experience. User Model User-Adapt Interact22.,101-123(2012)十字架谷歌学术
  10. 10.
    McNee,S.M.,Riedl,J.,Konstan,J.A。:准确是不够的:准确度指标有什么伤害推荐系统。:在计算系统中的人为因素会议 - 诉讼程序,PP。1097-1101(2006)谷歌学术
  11. 11.
    Tintarev, N., Masthoff, J.: Explaining recommendations: design and evaluation. In: Recommender Systems Handbook, pp 353–382. Springer, Boston (2015)谷歌学术
  12. 12.
    Herlocker,J.L.,Konstan,J.A.,Riedl,J.:解释协作过滤建议。在:ACM会议上的COMPERS COOP WORK WOP,PP。241-250(2000)谷歌学术
  13. 13。
    Swearingen, K., Sinha, R.: Interaction design for recommender systems. In: Designing Interactive Systems, pp. 1–10 (2002)谷歌学术
  14. 14。
    Gedikli,F.,Jannach,D.,GE,M .:我应该如何解释?建议系统的不同解释类型的比较。int j嗡嗡声计算螺柱72.,367-382(2014)十字架谷歌学术
  15. 15。
    Lamy,J.B.,Sekar,B.,Guezennec,G.等:解释了乳腺癌的人工智能:一种基于视觉案例的推理方法。ARIF INTERL MED.94,42-53(2019)十字架谷歌学术
  16. 16。
    Weitz,K.,Schiller,D.,Schlagowski,R.等:“让我解释!”:探索虚拟代理的潜力在解释的AI交互设计中。J多模用户界面(2020)。https://link.beplay登入springer.com/article/10.1007/S12193-020-00332-0_CITEAS.
  17. 17。
    Gilpin, L.H., Bau, D., Yuan, B.Z., et al.: Explaining explanations: an overview of interpretability of machine learning. In: Proceeding - 2018 IEEE 5th International Conference on Data Science and Advanced Analytics, DSAA 2018, pp. 80–89 (2019)谷歌学术
  18. 18。
    Bharadhwaj,H.,Joshi,S .:时间建议的解释。Ki - Kunstl Intellialz32.,267-272(2018)十字架谷歌学术
  19. 19。
    Ferreira,J.J.,Monteiro,M.S.:人们在Xai用户体验做什么?AI解释性研究与实践调查。在:Marcus,A.,Rosenzweig,E。(EDS。)设计,用户体验和可用性。当代交互环境设计。HCII 2020.计算机科学的讲义笔记,卷。12201. beplay登入Springer,Cham(2020)谷歌学术
  20. 20.
    Miller, T., Howe, P., Sonenberg, L.: Explainable AI: beware of inmates running the asylum or: how I learnt to stop worrying and love the social and behavioural sciences (2017)谷歌学术
  21. 21.
    BAAI: Beijing AI Principles (2019).https://www.baai.ac.cn/news/beijing-ai-principles-en.html.。Accessed 07 May 2020
  22. 22.。
    Ministry of Internal Affairs and Communications (MIC) the G of J.: AI R&D Principles (2017)谷歌学术
  23. 23。
    Google对:People + AI指南(2019)。https://pair.withgoogle.com/。Accessed 07 May 2020
  24. 24。
    Future of Life Institute (FLI): Asilomar AI Principles (2017).https://futureoflife.org/ai-principles/。Accessed 07 May 2020
  25. 25。
    微软:Microsoft AI原则(2018)。https://www.microsoft.com/en-us/ai/our-aproach-to -ai.。Accessed 07 May 2020
  26. 26。
    经合组织理事会关于人工智能的建议:AI(2019年)的经合组织原则。https://www.oecd.org/dour-digital/ai/principles/。Accessed 07 May 2020
  27. 27。
    European Group on Ethics in Science and New Technologies (EGE), European Commission Statement on artificial intelligence, robotics and “autonomous” systems : Brussels (2018)谷歌学术
  28. 28。
    Evenson,s .:术语讲故事:解释设计的经验。在:Bennett,A。(Ed。)设计研究,第231-240页。普林斯顿建筑出版社,纽约(2006年)谷歌学术
  29. 29。
    Budiu, R.: Can users control and understand a UI driven by machine learning? (2018).https://www.nngroup.com/articles/machine-learning-ux///。访问20月22日11月22日
  30. 30.
    Ahmad, M.A., Eckert, C., Teredesai, A.: Interpretable machine learning in healthcare. In: Proceedings of the 2018 ACM International Conference on Bioinformatics, Computational Biology, and Health Informatics, pp 559–560. ACM, New York (2018)谷歌学术

Copyright information

©作者,根据斯特普尔自然瑞士股份有限公司的独家许可beplay登入

作者和附属机构

  1. 1。LEUI, Laboratory of Ergodesign and Usability of InterfacesPUC-Rio里约热内卢巴西

Personalised recommendations