| 郭炜炜,王琦.人-无人车交互中的可解释性交互研究[J].包装工程,2020,41(18):22-28. |
| 人-无人车交互中的可解释性交互研究 |
| Explainable Interaction in Human-Autonomous Vehicle Interaction |
| 投稿时间：2020-07-29 修订日期：2020-09-20 |
| DOI：10.19554/j.cnki.1001-3563.2020.18.003 |
| 中文关键词: 人车共驾 人-无人车交互 可解释性人工智能 透明度 |
| 英文关键词: human-autonomous vehicle co-driving human-autonomous vehicle interaction explainable artificial in-telligence transparency |
| 基金项目:中央高校基本科研业务经费专项资金项目（22120190209） |
|
| 摘要点击次数: |
| 全文下载次数: |
| 中文摘要: |
| 目的 随着现代人工智能技术在自动驾驶系统中的广泛应用，其可解释性问题日益凸显，为此探讨人-无人车交互过程中的可解释性交互的框架以及设计要素等问题，以增强自动驾驶系统的决策透明性、安全性和用户信任度。方法 结合可解释人工智能和人机交互的基本理论与方法，本文首先介绍了可解释性人工智能，对当前可解释内容的提取方法进行总结，然后以人-机器人交互的透明度模型为基础，建立人-无人车交互中可解释性交互的框架。最后从解释的对象、方式和评价等多个设计维度对可解释性的交互设计问题进行探讨，并结合案例进行分析。结论 可解释性作为人与模型决策之间的接口，不仅仅是一个人工智能技术问题，而且与人密切相关，涉及到人-无人车交互中的多个层次。本文提出人-无人车交互中可解释性交互的框架，得出在人-无人车交互每个阶段需要的解释内容以及可解释交互设计的要素。 |
| 英文摘要: |
| With the wide application of modern artificial intelligence technology in autonomous driving system, its interpretability problem has become increasingly prominent. The work aims to discuss the explainability and interpretability of the autonomous systems during the interaction between the human and autonomous vehicles, in order to enhance the decision-making transparency, safety and user trust of the autonomous driving systems. Combining the fundamental theory and methodology of explainable artificial intelligence and human-computer interaction, explainable artificial intelligence (XAI) was firstly introduced and then the current methods and techniques of XAI generating the explanations were summarized. Based on the transparency model of human-robot interaction, the interaction framework for the explanations between human and autonomous vehicles was established. Finally, interaction design for explanations was discussed from the aspects of users, form and evaluation. As an interface between people and decision model, explainability is not only a domain of artificial intelligence but closely related to the people, involving multiple levels of the human-autonomous vehicle interactions (HAI). This paper proposes the interaction framework for explainability of HAI and obtains the contents requiring explanation in each stage of HAI and elements that can explain the interactive design |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
