Social XAI aims to co-construct an explanation jointly with an individual user in a specific situation. This requires mental models of the situation, the explanandum, and the interaction partner. These models must be dynamic in order to react to changes in the context and to new insights from the interaction. We discuss the characteristics and prerequisites of as well as the changes in these models against the background of human–human as well as human–computer and human–robot social interactions, and we present practical examples of how such models can be implemented in information systems using knowledge graphs.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Models of the Situation, the Explanandum, and the Interaction Partner

  • Heike M. Buhl,
  • Anna-Lisa Vollmer,
  • Rachid Alami,
  • Meisam Booshehri,
  • Kary Främling

摘要

Social XAI aims to co-construct an explanation jointly with an individual user in a specific situation. This requires mental models of the situation, the explanandum, and the interaction partner. These models must be dynamic in order to react to changes in the context and to new insights from the interaction. We discuss the characteristics and prerequisites of as well as the changes in these models against the background of human–human as well as human–computer and human–robot social interactions, and we present practical examples of how such models can be implemented in information systems using knowledge graphs.