IHAplanations on Demand - a Technique for hliciting the Actual Need for Explanations

2023 IEEE 31ST INTERNATIONAL REQUIREMENTS ENGINEERING CONFERENCE WORKSHOPS, REW(2023)

引用 0|浏览0
暂无评分
摘要
Explainability is an emerging quality aspect that is just beginning to be explored in the context of requirements engineering. Since there are many different types of explanations, it is not trivial to determine where users need which type of explanations. As the need for explanations depends on the user and the context, the user should participate in the elicitation process. However, previous research has shown that simply asking users if they want explanations does not lead to meaningful results, since users tend to answer this question in the affirmative, regardless of whether they need the explanation or not. To address this problem, we developed a technique that allows users to realistically put themselves in the context of the system and thus determine where explanations are really required. This technique requires a framework as requirements that are being raised with this approach relate to the different types of explanations. These types of explanations need to be identified and classified. To evaluate our technology, we conducted a user study with 21 participants. With the help of our technique supported by the framework, requirements for explainability can be elicited more realistically.
更多
查看译文
关键词
Explainability, Requirements Engineering, Elicitation techniques, Frameworks
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要