Data61's StellarGraph launches new algorithms for data analysis

StellarGraph — a division of CSIRO’s Data61 arm — has just launched three new algorithms to help researchers perform network graph analysis (NGA), and make better sense of large and complex datasets.

The new tools — an addition to its graph machine learning library — trump simpler algorithms, given their ability to spot data patterns easily across multiple sources.

By capturing data as a network graph, engineers and data scientists can, in turn, build and test powerful machine learning tools, using their own network data.

The firm’s existing toolkit has already been used to predict Alzheimer’s genes and detect Bitcoin ransomware, with ongoing applications being trialled in agtech.

StellarGraph Team Leader Tim Pitman says these new, more advanced, algorithms could open up further industry applications for the technology.

“The new algorithms in this release open up the library to new classes of problems to solve, including fraud detection and road traffic prediction,” he said.

“We’ve also made the library easier to use and worked to optimise performance allowing our users to work with larger data.”

Alex Collins, Group Leader Investigative Analytics, CSIRO’s Data61, added, “The challenge for organisations is to get the most value from their data. Using network graph analytics can open new ways to inform high-risk, high-impact decisions.”

StellarGraph said the tools will prove particularly useful in fields like cybercrime analysis, which use highly complex and densely connected data from multiple sources.

However, further applications could also be seen in areas such as contact tracing for infectious disease.

The algorithms are now freely available to the open source community on GitHub.

Image credit: ©stock.adobe.com/au/adiruchnachiangmai