Researcher Access Protocols

LearnAdapt is deeply committed to the principles of Open Science. We believe that collaborative, peer-reviewed research is essential to building ethical and effective educational AI. To facilitate this, we provide structured access to our anonymized datasets and core architectural models for academic researchers, data scientists, and educational institutions.

1. Dataset Collaboration (Anonymized Telemetry)

To catalyze research in Educational Data Mining (EDM), we provide access to de-identified datasets containing high-fidelity learning trajectories, cognitive traces, and model interaction logs.

Prerequisites for Access

• Institutional Review Board (IRB) Approval: You must provide proof of IRB approval (or exemption) from your home institution for your specific research protocol.
• Data Use Agreement (DUA): Lead researchers must sign a DUA stipulating that no attempts will be made to re-identify users or institutions. Data must not be shared outside the approved research team.
• Academic Affiliation: Primary applicants must be affiliated with an accredited academic or research institution.

Request Process

Submit an application to research@learnadaptresearch.org including:

1. Principal Investigator CV.
2. IRB approval documentation.
3. A brief research proposal (max 2 pages) outlining your hypotheses, methodology, and how the LearnAdapt dataset will be utilized.

2. Testing Architectural Assumptions (Sandbox Access)

For researchers aiming to evaluate our cognitive architecture, prompt-routing heuristics, or Bayesian Knowledge Tracing (BKT) implementations, we offer dedicated Sandbox access.

Sandbox Environment Capabilities

• Isolated Evaluation: A cloned, containerized instance of the LearnAdapt backend, separate from production data.
• API Interoperability: Access to our proprietary orchestration APIs, allowing you to inject synthetic learner states and measure system responses.
• Model Switching: Ability to benchmark our default routing heuristics against your own custom LLM wrappers or classification models.

Request Process

Submit a technical proposal to research@learnadaptresearch.org detailing:

1. The specific architectural components you intend to evaluate.
2. Your technical methodology (e.g., A/B testing, latency benchmarking, predictive accuracy analysis).
3. Expected computational resource requirements.

3. Publication and Co-Authorship Guidelines

We actively encourage the publication of empirical findings derived from the LearnAdapt platform.

• Citation Requirement: All resulting publications must appropriately cite the foundational LearnAdapt architectural manuscript and relevant dataset DOIs.
• Pre-Publication Review: We request a 14-day pre-publication review window. This review is strictly limited to ensuring the accurate technical representation of our architecture and verifying that privacy protocols were maintained. We do not exercise editorial control over your empirical findings.
• Co-Authorship: We welcome collaborative research proposals where our internal data science team can contribute to the methodological design and analysis, warranting co-authorship.

We look forward to collaborating with you to advance the frontier of transparent, adaptive learning science.