AQUAGEN.AI is a new software product developed by The Learning Science Group that utilizes advanced AI prompt engineering techniques to generate exam questions for the Navy. Unlike previous automatic question generation systems, AQUAGEN.AI will use Large Language Models (LLMs) trained on extensive text data and a modular architecture to future-proof the system.
AQUAGEN.AI's technique of layering AI models involves combining the outputs of multiple models into a single, more accurate prediction. This allows the system to leverage the strengths of each individual model for better performance than using a single model alone. Additionally, the modular architecture allows for the best AI models to be used based on requirements, such as security, and allows new AI models to be easily incorporated as AI technologies advance.
The system is designed to conform to NETC-N7 guidance and incorporate the Practical Applications of Learning Science (PALS) principles, which promote acquisition, long-term retention, and transfer of training. AQUAGEN.AI will integrate the PALS principles of retrieval practice, interleaving, spaced practice, and elaboration to enhance acquisition, long-term retention, and transfer of training. The system will generate questions not only for in-class exams but also for in-class knowledge checks and practice exams for self-study.
To ensure the quality of exam questions and instructional materials, AQUAGEN.AI will include built-in tools for the psychometric analysis of exam responses.
Automated psychometric analyses in AQUAGEN.AI will include Item Difficulty Index, Discrimination Index, Point-Biserial Correlation Coefficient, and the Kuder-Richardson Formula 20 (KR-20). Instructors and designers will be provided with easy-to-understand reports along with recommendations for action.
AQUAGEN.AI will also include automated content analysis and content alignment. Content analysis identifies key concepts and knowledge areas taught in the course, while content alignment verifies the accuracy and relevance of each exam question. The former will enable instructional designers to produce new instructional materials more quickly than they currently can by mining the knowledge base generated by AQUAGEN.AI for key concepts aligned with instructional objectives. The latter can be used to provide remediation to learners on missed exam questions. Correct responses, as well as incorrect responses, are directly linked to the source information in the knowledge base allowing for quick lookups for contextual clarification.
Overall, AQUAGEN.AI promises to revolutionize the way exam questions are generated and improve the quality of instructional materials. The system's use of advanced AI prompt engineering techniques and Large Language Models trained on extensive text data, modular architecture, and incorporation of PALS principles ensures the system is effective, efficient, and future-proofed. The built-in tools for the psychometric analysis of exam responses and automated content analysis and alignment will provide instructors and designers with easy-to-understand reports along with recommendations for action. With AQUAGEN.AI, generating exam questions and producing instructional materials will be faster, easier, and more effective.