Learning to Delete: Forgetting of Digital Objects as Collaborative Task of Human and AI

学习删除：忘记数字对象作为人类和人工智能的协作任务

• 批准号: 427404493
• 负责人: Professorin Dr. Cornelia Niessen
• 金额: --
• 依托单位: Lehrstuhl für Psychologie im Arbeitsleben
• 依托单位国家: 德国
• 项目类别: Priority Programmes
• 财政年份: 2019
• 资助国家: 德国
• 起止时间: 2018-12-31 至 2023-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/427404493?language=en
• 关键词: Learning; Delete; Forgetting; Digital; Objects

In the digital world of work, data can in principle be stored in unlimited quantities. Unlike human memory, external memory does not forget. Digital objects that have become irrelevant can hinder the search for information, delay decision-making processes, divert attention away from the actual task and thus impair work performance and well-being. In order to be able to work in a focused and goal-oriented way, forgetting processes in the joint cognitive system become increasingly important. People not only have to ignore irrelevant information and deliberately inhibt it in their memory, but also "forget" it in external memory by hiding, archiving or permanently deleting this information. This is where the cognitive companion Dare2Del comes into play, whose basic functions based on inductive logic programming (ILP) were developed in the first application period on the basis of cognitive and occupational psychology research on intentional forgetting in the working context (field 1 and 2 of the SPP). The goal of Dare2Del is to support employees in administration and production to regulate their digital knowledge by hiding and deleting irrelevant digital objects. Dare2Del is designed as an interactive, intelligent assistance system that adapts to users through incremental learning. In the first application period focus was on the need-based support of short-term forgetting processes by fading out irrelevant information; in the second application period the work focuses on the function of deletion. People often store too many irrelevant digital objects in an unstructured way, rarely delete them, often doing so only when outdated information interferes with the workflow. Deletion is accompanied by uncertainty and is often neglected in everyday working life. Consequently, the deletion recommendations of the forgetting information system Dare2Del should be designed transparently and comprehensible such that people actually deal with these decisions and that performance and stress level are positively influenced. This design task addresses work area 4 of the SPP. It is planed to empirically study which design features of the cognitive companion, which personal (e.g., ability to inhibit) and organizational conditions (e.g., social norms), do support employees in their decision making about which digital objects can be deleted. The companion system Dare2Del is to be extended by a component for the generation of different types of explanations. A collaborative learning approach will be developed, allowing users to mark portions of the generated statements that they do not accept which are then considered as constraints in incremental learning.

在工作的数字世界中，数据原则上可以无限量地存储。与人类记忆不同，外部记忆不会忘记。变得无关紧要的数字对象可能会阻碍信息搜索，延迟决策过程，转移对实际任务的注意力，从而损害工作绩效和福祉。为了能够以专注和目标导向的方式工作，联合认知系统中的遗忘过程变得越来越重要。人们不仅要忽略不相关的信息，故意将其隐藏在记忆中，还要通过隐藏、存档或永久删除这些信息，将其“遗忘”在外部记忆中。这就是认知伴侣Dare2Del发挥作用的地方，其基于归纳逻辑编程（ILP）的基本功能是在第一个应用阶段基于对工作环境中有意遗忘的认知和职业心理学研究（SPP的领域1和2）开发的。Dare2Del的目标是通过隐藏和删除不相关的数字对象来支持管理和生产员工管理他们的数字知识。Dare2Del是一个交互式的智能辅助系统，通过增量学习来适应用户。在第一个应用期间的重点是基于需要的支持短期遗忘过程中淡出不相关的信息，在第二个应用期间的工作重点是删除的功能。人们经常以非结构化的方式存储太多不相关的数字对象，很少删除它们，通常只有在过时的信息干扰工作流程时才这样做。删除伴随着不确定性，在日常工作生活中往往被忽视。因此，遗忘信息系统Dare2Del的删除建议应该设计得透明和易于理解，这样人们就可以真正处理这些决定，并且对性能和压力水平产生积极影响。本设计任务涉及SPP的工作领域4。计划实证研究认知伴侣的哪些设计特征，哪些个人（例如，抑制能力）和组织条件（例如，社会规范），确实支持员工做出关于可以删除哪些数字对象的决策。配套系统Dare2Del将由一个组件扩展，用于生成不同类型的解释。将开发一种协作学习方法，允许用户标记他们不接受的生成语句部分，然后将其视为增量学习中的限制。