Creating a NASA Data Oriented Transformer Framework for Next Generation Retrieval and Modeling

We seek to investigate whether the transformer/foundation model framework can overcome two common hurdles when applying ML/AI techniques to NASA data, which are (1) cross-time and cross-mission knowledge transfer, and (2) physics-consistency and transparency. This project will build an ABI pretrained transformer/foundation model, and will be used for a 3D cloud reconstruction application to determine the benefits.

A pretrained model using the ABI observations will be built integrating both ABI's spatial and temporal information. This model will be published in the public domain, which can be used for general purposes using ABI data. Meanwhile we will fine-tune the model to a 3D cloud mask prediction case to evaluate (1) the benefit of improving the prediction performance; (2) the computational benefit from transferred learning. An explainability metrics will also be developed out of this project to evaluate the physics-consistency of model decision process.

This is early/mid-stage (TRL 3) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.