Connect with us
Follow us on social media
From online shopping and streaming music to map directions and video games, there’s a mobile app for just about everything.
In fact, as of January 2018, mobile phone users had the choice of some 3.5 million apps via the Google Play Store and 2.1 million apps from Apple – the two leading app stores in the world.
With mobile apps downloads globally expected to exceed 352.9 billion by 2021, the need for automated vetting of apps at entry is becoming increasingly important to prevent fake apps, possibly containing malware, from entering the marketplace.
“Entering credit card information opens a customer to potential financial fraud,” says Dr Suranga Seneviratne from the Centre for Distributed and High-Performance Computing (CDHPC).
“Some fake apps contain malware that can steal personal information or even lock the phone until the user pays a ransom. Others encourage users to log in using Facebook credentials, potentially exposing sensitive personal information.
“Automated vetting of apps at entry is becoming a must and these methods have to include various checkpoints to detect different types of malware and dubious practices followed by some app developers.”
To improve vetting practices, Dr Seneviratne and his colleagues are researching deep learning models capable of detecting counterfeit apps in the mobile device market.
The project is being supported by the Google Faculty Research Awards program, which is designed to recognise and support world-class, permanent faculty staff pursuing cutting-edge research in areas of mutual interest.
The award is highly competitive with only 15% of applicants receiving funding, and each proposal undergoing a rigorous Google-wide review process.
Dr Seneviratne’s project was one of only 152 projects accepted in the latest round of entries, from an original pool of 1,033 research proposals from more than 360 universities across 46 countries.
“We are very happy to receive this award. It will enable us to conduct cutting-edge research in mobile security, an aspect of cybersecurity that is crucial for billions of smartphone users,” says Dr Seneviratne.
“We intend to focus on developing deep learning models that are capable of detecting counterfeit apps in mobile app markets.
“Our aim is to enable consumers to freely choose apps without the fear of malware or identity theft, and the associated costs to them and the community.”