A Cognitive Assistant for improving human reasoning skills. Issue 117 (September 2018)

• Record Type: Journal Article
• Authors: Le, Nguyen-Thinh; Wartschinski, Laura
• Is Part Of: International journal of human-computer studies. Issue 117(2018)
• Page Start: 45
• Page End: 54
• Publication Date: 2018-09
• Subjects: Dialog-based systems -- Human reasoning -- Evaluation -- Heuristics and biases -- Cognitive Assistant; Human-machine systems -- Periodicals; Systems engineering -- Periodicals; Human engineering -- Periodicals; Human engineering; Human-machine systems; Systems engineering; Periodicals; Electronic journals; 004.019
• DOI: 10.1016/j.ijhcs.2018.02.005
• Languages: English
• ISSNs: 1071-5819

Abstract

Highlights: In this work, a cognitive assistant was presented that was designed to improve skills for a selection of core reasoning tasks. Although it did not have astonishing success in all the tasks, maybe because Wason's Selection Task and Bayesian reasoning are both very complicated concepts to teach, it did show very satisfying results with the other tasks. The cognitive assistant was clearly more effective than an online course on the same skills, proving its worth as a mentor. Especially having in mind that participants only spent roughly 30 min talking to it, it can clearly be viewed as a success. This article contributes to the community of cognitive assistants two-fold: (1) a dialog-based cognitive assistant for improving human reasoning skills, which is able to adapt to individual performance of the student has been proposed and realized; (2) Over sixty different reasoning tasks and their necessary data (complete with explanations, associated small talk, hints, evaluations etc.) adopted from research in psychology have been integrated in the cognitive assistant, which has been empirically evaluated and confirmed that the cognitive assistant yields higher learning gains than a comparative non-interactive environment. Abstract: Cognitive Assistants support humans and enhance their capabilities in solving a wide variety of complex tasks. In this paper, we introduce a Cognitive Assistant named LIZA that is developed as a pedagogical agent and aims at improving the reasoning and decision making abilities of its users. This Cognitive Assistant is able to hold conversation with users in natural language in order to help them solve problems of common heuristics and biases. Using controlled experiments, we demonstrate that LIZA could help test persons improve their reasoning skills, and show that LIZA achieved significant higher learning gains than a non-interactive online course.