Personalized Education with Generative AI and Digital Twins: VR, RAG, and Zero-Shot Sentiment Analysis for Industry 4.0 Workforce Development

TL;DR

The paper addresses the need for scalable, personalized 4IR workforce training amid widespread automation and gaps in access for underrepresented groups. It introduces gAI-PT4I4, a Generative AI–based Personalized Tutor that integrates a Multi-Fidelity Digital Twin Education Framework, VR-enabled learning interfaces, zero-shot sentiment analysis, and retrieval-augmented generation (RAG) to tailor Industry 4.0 coursework. Key contributions include a reference model mapping digital twin fidelities to Bloom's Taxonomy and Kirkpatrick evaluation, a Unity-based VR learning interface with an Interactive Instructor, a zero-shot sentiment analysis pipeline validated on EduTalk and TSATC datasets achieving $0.86$ accuracy, and GraphRAG-enhanced LLMs for domain-grounded tutoring, all validated with 22 volunteers showing improved learning metrics and reduced training time. The framework offers a scalable, domain-grounded approach to 4IR workforce development, with potential to boost URM retention and engineering identity through adaptive, data-informed instruction.

Abstract

The Fourth Industrial Revolution (4IR) technologies, such as cloud computing, machine learning, and AI, have improved productivity but introduced challenges in workforce training and reskilling. This is critical given existing workforce shortages, especially in marginalized communities like Underrepresented Minorities (URM), who often lack access to quality education. Addressing these challenges, this research presents gAI-PT4I4, a Generative AI-based Personalized Tutor for Industrial 4.0, designed to personalize 4IR experiential learning. gAI-PT4I4 employs sentiment analysis to assess student comprehension, leveraging generative AI and finite automaton to tailor learning experiences. The framework integrates low-fidelity Digital Twins for VR-based training, featuring an Interactive Tutor - a generative AI assistant providing real-time guidance via audio and text. It uses zero-shot sentiment analysis with LLMs and prompt engineering, achieving 86\% accuracy in classifying student-teacher interactions as positive or negative. Additionally, retrieval-augmented generation (RAG) enables personalized learning content grounded in domain-specific knowledge. To adapt training dynamically, finite automaton structures exercises into states of increasing difficulty, requiring 80\% task-performance accuracy for progression. Experimental evaluation with 22 volunteers showed improved accuracy exceeding 80\%, reducing training time. Finally, this paper introduces a Multi-Fidelity Digital Twin model, aligning Digital Twin complexity with Bloom's Taxonomy and Kirkpatrick's model, providing a scalable educational framework.

Figures (7)

• Figure 1: Multi-Fidelity Digital Twin Education Reference Model
• Figure 2: Generative AI-based Personalized Tutor for Industrial 4.0 (gAI-PT4I4) Framework
• Figure 3: Zero-shot LLM-based sentiment analysis pipeline
• Figure 4: Learning Interface built with Unity
• Figure 5: Zero-shot Qualitative to Quantitative Sentiment Analysis in Teacher-Student Conversations (Using Table \ref{['tab:TS_Conv_Case']} as Case Analysis)