Marine - eLearning Design and Guidance Engine (Marine-EDGE)

US Marine Corps (USMC) training and education (T&E) programs are critical to ensure mission readiness for executing expeditionary operations. Current T&E programs are not adequately preparing the USMC for the future operating environment. Many current T&E programs rely on outdated programs of instruction (POIs) that use static slides, written exams, and minimal experiential learning. Recent advancement of generative artificial intelligence (GenAI) technologies provide an opportunity to transform current USMC T&E systems, making it easier to design and implement targeted course materials, develop more effective courses, and accelerate student learning. The USMC seeks to modernize its T&E by integrating AI-enabled tools that support instructional system design, content generation, and legacy content conversion. However, the current instructional landscape presents critical structural, technical, and cultural challenges. POI curriculum design and approval is manual and time consuming. Legacy course content is composed of outdated static materials and minimal experiential learning. Course content generation is complex and challenging to tailor to student and instructor needs at scale.To meet the above challenges and provide the desired capabilities, Machina Cognita Technologies (MCT) and Cognitive Performance Group (CPG) propose the Marine – eLearning Design and Guidance Engine (Marine-EDGE). Marine-EDGE is a USMC instructional design and content generation assistant that supports curriculum developers, POI managers, and instructors throughout course planning and with real-time content delivery. It provides guidance at each stage of the instructional design process to ensure alignment with training objectives, learning science principles, and operational needs. A natural language processing (NLP) knowledge graph (KG) will be constructed and learn representations of exemplary course design through CPG’s deep expertise in human-centered design (HCD) approaches and cognitive science. A fine-tuned Mixture of Experts (MoE) large language model (LLM) with graph-retrieval augmented generation (RAG) will be trained to rapidly generate targeted course and curriculum guidance as a design assistant to POI course directors. A curriculum design pipeline, supported by a plug-and-play LLM architecture, will accelerate USMC POI development and approval through interactive curriculum design and expert course curation advice. Finally, a content recommendation pipeline will facilitate rapid prototyping of updated interactive content for both course creation and legacy content conversion. Across all POI development workflows, personnel will interact with GenAI assistance through an intuitive graphic user interface (GUI) that supports user selected templates, drag-and-drop content, and conversational chatbot messaging. Altogether, the system will modernize USMC eLearning, increase Marine schoolhouse production, and improve operational mission readiness.