TY  - THES
AB  - Language education is a global and important issue nowadays, especially for young children since their later educational success build on it. But learning a language is a complex task that is known to work best in a social interaction and, thus, personalized sessions tailored to the individual knowledge and needs of each child are needed to allow for teachers to optimally support them. However, this is often costly regarding time and personnel resources, which is one reasons why research of the past decades investigated the benefits of Intelligent Tutoring Systems (ITSs). But although ITSs can help out to provide individualized one-on-one tutoring interactions, they often lack of social support.<br />
This dissertation provides new insights on how a Socially Assistive Robot (SAR) can be employed as a part of an ITS, building a so-called "Socially Assistive Robot Tutoring System" (SARTS), to provide social support as well as to personalize and scaffold foreign language learning for young children in the age of 4-6 years. As basis for the SARTS a novel approach called A-BKT is presented, which allows to autonomously adapt the tutoring interaction to the children's individual knowledge and needs. The corresponding evaluation studies show that the A-BKT model can significantly increase student's learning gains and maintain a higher engagement during the tutoring interaction. This is partly due to the models ability to simulate the influences of potential actions on all dimensions of the learning interaction, i.e., the children's learning progress (cognitive learning), affective state, engagement (affective learning) and believed knowledge acquisition (perceived learning). This is particularly important since all dimensions are strongly interconnected and influence each other, for example, a low engagement can cause bad learning results although the learner is already quite proficient. However, this also yields the necessity to not only focus on the learner's cognitive learning but to equally support all dimensions with appropriate scaffolding actions. Therefore an extensive literature review, observational video recordings and expert interviews were conducted to find appropriate actions applicable for a SARTS to support each learning dimension. The subsequent evaluation study confirms that the developed scaffolding techniques are able to support young children’s learning process either by re-engaging them or by providing transparency to support their perception of the learning process and to reduce uncertainty. Finally, based on educated guesses derived from the previous studies, all identified strategies are integrated into the A-BKT model. The resulting model called ProTM is evaluated by simulating different learner types, which highlight its ability to autonomously adapt the tutoring interactions based on the learner's answers and provided dis-engagement cues. Summarized, this dissertation yields new insights into the field of SARTS to provide personalized foreign language learning interactions for young children, while also rising new important questions to be studied in the future.
DA  - 2019
DO  - 10.4119/unibi/2939409
LA  - eng
PY  - 2019
TI  - Integrating Socially Assistive Robots into Language Tutoring Systems. A Computational Model for Scaffolding Young Children's Foreign Language Learning
UR  - https://nbn-resolving.org/urn:nbn:de:0070-pub-29394099
Y2  - 2024-11-22T05:25:14
ER  -