human/robot mutual learning

The research is a collaboration between the DICE-lab at University of Gothenburg (GU) and the ETIS lab at Cergy CY Université which was initiated via EUTOPIA’s Undergraduate Research Support Scheme EURSS which was granted to Alva Markelius and Sofia Sjöberg spring 2022. The project was developed by the combination and collaboration of the GU expertise in Differential Outcomes Training (DOT) and human robot interaction, and the Cergy CY expertise in robot language development and robot/caregiver interaction. In June 2022 we visited the ETIS lab at Cergy to develop the mutual learning interaction setup with the robot Reachy (Pollen robotics). In August later that year the setup was piloted with 6 participants in Sweden.

The purpose of the research is to improve and gain insight in how DOT can be implemented as a component of robot language acquisition in a collaborative interactive setup for robot and human homeostatic mutual learning. The setup empowers humans for training their discriminative working memory while at the same time interactively teaching a social robot language through affective interactions. The robot used in the study has a vocabulary similar to that of a human infant in the linguistic “babbling” phase. The robot software architecture is built upon a model for affect-grounded language acquisition, where the robot associates vocabulary with internal needs (hunger, thirst, curiosity) through interactions with the human.

The resulting work is reported in the paper “A Human-Robot Mutual Learning System with Affect-Grounded Language Acquisition and Differential Outcomes Training”. We chose to adopt a differential outcomes training (DOT) protocol with the aim of facilitating the interactive learning and found evidence that DOT can enhance the human’s learning efficiency, which in turn enables more efficient robot language acquisition. The results of the initial pilot study revealed that the robot’s language acquisition achieves higher convergence rate in the DOT condition compared to the non-DOT control condition. Additionally, participants reported positive affective experiences, feeling of being in control, and an empathetic connection with the robot. This mutual learning collaborative approach offers a potential contribution to improve cognitive interventions with DOT (e.g. for people with dementia) and treatment adherence through helping humans stay more engaged by taking a more active role in the interaction. The homeostatic motivational grounding of the robot’s language acquisition has potential to contribute to more ecologically valid and socially appropriate social interactions.