Teachy Mini: Knowledge-Based Generative Social Robot for Higher Education Tutoring
Researchers operationalized knowledge-based design requirements in the Reachy Mini robot platform to create Teachy Mini, a generative social robot tutoring system for higher education. The system uses system prompting, retrieval-augmented generation, and stateful prompt orchestration. A preliminary evaluation study with 24 participants compared robot-guided learning sessions about research methodologies using Teachy Mini versus a control version. The study aims to address risks of misinformation and lack of transparency in large language model-powered educational robots.
Key facts
- Teachy Mini is a generative social robot (GSR) tutoring system for higher education.
- It is built on the Reachy Mini robot platform.
- The system uses knowledge-based design (KBD) requirements.
- Techniques include system prompting, retrieval-augmented generation, and stateful prompt orchestration.
- A preliminary evaluation study involved 24 participants.
- Participants completed a robot-guided learning session about research methodologies.
- The study compared Teachy Mini with a control version.
- The work addresses risks of misinformation and lack of transparency in LLM-powered robots.
Entities
Institutions
- Teachy Mini
- Reachy Mini