AI-ALOE Technology Infrastructure Architecture

The horizontal layers of the architecture represent the flow of information from the low-level AI agents to the user dashboard. All the layers shall have governance and oversight mechanisms and work together to serve AI-ALOE stakeholders.

At the bottom of the conceptual architecture diagram are AI services such as Jill Watson question answering (Q/A) agent, Apprentice tutors, SMART, SAMI, AskJill self-explanation agent, VERA, and embedded video tutors. Together these AI services offer teacher presence, cognitive presence, self-directed learning, social presence, and self-explanation.

Data from teachers, learners, and researchers flow from the AI agents, Learning Management Systems, online discussion forums, etc. flow into the local data management layer. In this layer,  the core research teams store, anonymize, and analyze the data locally.

Next, data is uploaded to the AI-ALOE data management layer. In this layer, data is ingested, stored, parsed, organized, anonymized, standardized, and is finally available for search, download, and cataloging. Users such as learners, teachers, and researchers can access (via an access API) the available data, built-in data analytics (performance metrics, leaner trace, interventions, etc.), and helpful visualizations (learning patterns, performance, challenges, network, and feedback) through the user dashboard.

Overall, this architecture aims to improve learning outcomes through foundational AI focused on personalized learning, human-AI interaction, human-data interaction, data security and privacy, machine teaching, and self-directed learning. This is represented as the user impact in the top layer.