Researchers use analytics and machine learning to detect dodgiest travellers

Unisys and CSIRO’s Data61 are researching how data analytics could detect potential border security risks posed by passengers, visa applicants, cargo and parcels.

The research – funded by Unisys – involves assessing anonymised data sets from airlines using analytics and machine learning to identify patterns and flag risks including malicious traveller intent and dangerous cargo contents.

A proof-of-concept will be conducted at an unnamed “major Asian hub” with the intention to develop the technology into a product that will be sold to governments around the world, the organisations said.

The Australian Department of Immigration and Border Protection is already a client of Unisys, which is responsible for the department’s new border clearance platform.

“The end goal of this international collaboration is to make border security processes more efficient, cost effective and safer for countries around the world,” said Data 61 CEO, Adrian Turner.

“It’s one of the ways Data61 is working with industry to translate data-science – in this case deep analytics and machine learning – into a viable product to help deliver economic and societal impact.”

John Kendall, global border security director at Unisys, said the solution would help governments better focus resources on riskier travellers and cargo.

“Most border agencies today rely on human designed rules to identify suspicious people or cargo. Working with Data61, we are incorporating machine learning and real-time data analytics to reveal the actual intent of travellers and shippers,” he said.

“This will allow border agencies to automate the processing of low risk people and cargo while reserving specialised border security resources for the small percentage of travellers and cargo that present a higher risk profile.”