By David McNew/CBS News “It’s a project based learning project.”
That’s what the name implies.
And it’s the brainchild of researchers at the University of Illinois at Urbana-Champaign.
The University of Southern California is partnering with the University at Buffalo to develop a learning platform for data scientists that uses machine learning to create data visualizations, called a “visual analytics platform.”
The first version of the platform was published in November, and the researchers say they hope to eventually be able to leverage machine learning in the form of a database of visualizations for a broad range of topics.
“What the platform does is it builds on top of existing tools, and then it takes that and applies the algorithms,” said Alex Rios, an associate professor of computer science and director of the Computational Neuroscience Laboratory at U-S.C. “It then turns it into a visual database of data and you can use that to learn from the data.”
What makes this new technology particularly exciting is that it allows data scientists to use algorithms that have already been used to develop powerful visualizations.
“We can train an algorithm to learn by seeing how well it can predict a data set,” Rios said.
“In other words, we can learn how to predict the distribution of data sets.”
What the researchers hope to do with the data is to build a tool that can be used for more than just analyzing data.
They’re building a database for all sorts of applications.
For example, if you want to understand the effects of climate change, you can build a database and then use the database to predict when it will be warmer or colder in the future.
In this way, you could predict the future temperature of the future and then set up weather stations in the vicinity of those future locations.
And, if the researchers want to create visualizations that can predict future weather events like hurricanes, they could use their data to create forecasts of how that event might unfold.
“This is the sort of thing where the data can help us to make better decisions,” Roes said.
The project is currently in a closed beta phase.
There’s no timeline for when it’ll be publicly released.
But the researchers expect to begin training their AI to make predictions in the next two years.
Rios hopes that the new platform can help make the big data data that scientists collect more useful and valuable for the researchers they work with.
“If you can train your machine learning system to do these sorts of tasks, it opens up a lot of other things that are useful to us,” Rias said.
“The question is, is this technology going to work?
And we think that it’s going to be very important for this sort of work to be done in the context of big data,” he added.