Visual AI pipeline recommender for dataset-model matching
This technology creates a system that maps AI datasets and the computational pipelines used to process them into separate visual 'landscapes,' then models how well different pipelines work with different datasets. Think of it like a GPS-style terrain map where latitude and longitude show which datasets and pipelines are similar to each other, and elevation shows how well a given dataset-pipeline pairing actually performs. The system can then recommend which pipeline is the best fit for a new dataset by finding its location on the map and seeing what works well nearby. The goal is to help data scientists and ML engineers quickly navigate a large library of datasets and processing workflows without manually testing every combination.
What you could build
A visual analytics tool embedded in ML platforms or enterprise data science environments that recommends optimal processing pipelines for new datasets and surfaces unexplored dataset-pipeline pairings; primary buyers would be enterprise AI/ML platform vendors and large data science teams in finance, healthcare, or government.
Who in Virginia should care
Northern Virginia's dense concentration of federal data science contractors, defense analytics firms, and cloud infrastructure companies would be natural early adopters or licensees.
Readiness: Concept
Concept — described but not yet demonstrated. Lab validated — supported by experimental results in the patent. Prototype likely — the text describes a built, working embodiment.
Readiness is inferred from the patent text, not from a lab visit.
The record
- Inventors
- RAN JIN, Xiaoyu Chen, Seyedeh Parshin Shojaee
- Filed
- Patent pending — filed September 29, 2023
- Status
- Application
- Publication number
- US20240127038A1
Ready to talk?
Virginia Tech Intellectual Properties handles licensing for this technology.
Prosim summaries are generated from public patent text and are not legal advice.