Exploring Q-Learning in Social Robots for English-Slovak Vocabulary Learning

The study explores the use of social robots in the context of teaching new vocabulary to students. We employ reinforcement learning, specifically Q-learning, as our base algorithm for selecting which area of words should be learned by each student individually, so the students can improve areas where they have gaps. Experiments were designed to compare students with a personalized selection of words to learn versus those without it. This study incorporated words from both English and Slovak languages. We prove that the use of Q-learning has a positive impact on learning outcomes with an observed improvement of average grade points. We have also interviewed students and teachers to discover their thoughts about using the NAO robot as an assistant for vocabulary learning.