about mosaiks

Right now, numerous public satellite systems collect huge amounts of data about the world every day. But there is so much imagery (terabytes per day) that it’s overwhelming to sort through by hand; and it’s too complex and unstructured to be usable in its raw form for most applications.

That is why linking satellite imaging to machine learning (SIML) is incredibly powerful. It enables vast amounts of unstructured image data to be transformed into structured information that can immediately be used for planning, research, and decision-making.

We believe that people all over the world should be able to access SIML technologies, but we also recognize that most people who would benefit the most from these tools don’t have the time or resources to manage enormous satellite imagery data sets and learn how to apply machine learning to them.

That’s why we developed MOSAIKS (Multi-task Observation using SAtellite Imagery & Kitchen Sinks).

MOSAIKS is designed to work “out of the box” for a wide array of SIML applications, for people with no SIML expertise who work on normal desktop computers. MOSAIKS users never actually have to touch satellite imagery themselves and only need to have basic statistical training.

If you can run a regression, you can use MOSAIKS.

MOSAIKS empowers users to create their own new data sets from satellite imagery. We don’t control what variables users look at, and we never need to know. MOSAIKS is a system that allows users to quickly transform vast amounts of imagery into maps of new variables, using their own training data.

If you’ve ever been curious about trying machine learning with satellite imagery, but don’t know anything about machine learning or satellite imagery, MOSAIKS is for you.

And if you know a lot about machine learning and satellite imagery, MOSAIKS might still be for you, since it performs competitively with deep learning methods but is much simpler and cheaper to use.