“A humanitarian application of machine learning”

Ukraine is by some estimates now the most mine-contaminated country in the world and mines will continue to endanger lives even when the war is over.

“MineFree” is a mobile app that allows volunteers to alert the State Emergency Service of Ukraine to the presence of unexploded ordnance such as undetonated munition and land mines.

Volunteers take photographs of unexploded ordnance and send these via the app to the relevant authorities, who can then investigate.

Josh Freeman, a visiting Computer Science student last year at Oriel College, has been developing state-of-the-art machine learning technology to identify the devices captured by photographs and sort them in order of priority.

The technology would ensure that demining resources (which are limited) are directed to the most dangerous sites first, potentially saving lives.

“For me what the project signifies is that machine learning is not just about machines; it’s also about people. It’s a humanitarian application of machine learning,” Josh said.

Josh began working on the project at the frontend of his year studying at Oriel College when he received a message from a Ukrainian PhD student at his home institution, the Swiss Federal Institute of Technology Lausanne, asking for assistance on a machine learning programme aimed at enhancing demining operations in Ukraine.

Previously, Josh had helped to establish a society for Ukrainians at the university.

“It all just felt, just feels, really unjust,” he said, referring to Russia’s military invasion of Ukraine.

The year that Josh spent at Oriel College was the final year of a BSc in Computer Science degree course. He has now began studying for a master’s degree at ETH Zürich.

His academic interests lie in the field of quantum computing and while based at Oriel College he completed a research project with the Quantum Group, working with Professor Jonathan Barrett.