A study to improve social media data analysis, led by The University of Western Australia, has received a $150,000 grant to better understand defence analytics crucial to national security.

The study, Empowering the open-source intelligence (OSINT) analyst through explainable models, is one of three UWA collaborations awarded Defence Science Centre grants.

Project leader Dr Mehwish Nasim, from UWA's School of Physics, Mathematics and Computing, said sophisticated analytical models were vital to interpreting – and even predicting – the outcomes of high-stakes defence scenarios such as national security threats.

Image: Dr Mehwish Nasim.

“The ability to interpret complex algorithms ensures that decision-makers in defence can understand the logic behind the solution and make the most of human-machine teaming to reach more informed decisions,” Dr Nasim said.

Associate Professor Michael Wise and Dr Nasim are co-leading the study and the team has been working to provide greater transparency and deeper insight into open-source data particularly social media data, to develop explainable models.

“Transparency is critical to defence analytics, where decisions can have profound consequences,” Dr Nasim said.

“Explainable models play a vital role in providing defence analysts with clear insights into how algorithms arrive at their conclusions.”

Dr Nasim said the multidisciplinary collaboration with technology company PHASE and Edith Cowan University, represented by Dr Stephanie Meek, was expected to achieve a breakthrough in data analysis.

“The Defence Science Centre grant will allow the project to explore state-of-the-art techniques to improve defence analysts’ ability to interpret data, with major potential for real-world applications,” Dr Nasim said.