M’sian wins Aussie award for research

A PROPOSAL for an intelligent healthcare system safe and secure from cyberattacks has earned a Malaysian information technology (IT) student an award in Australia.

Arawinkumaar Selvakkumar (pic) from Subang Jaya, Selangor, received the Best Paper Awards from the Queensland University of Technology (QUT) School of Computer Science for his research paper titled “Addressing the Challenges and Impact of Adversarial Machine Learning Attacks on Smart Healthcare Systems”.

The 23-year-old said the outbreak of the Covid-19 pandemic sparked his interest in the healthcare industry and how it can adapt to meet the overwhelming demand for its services.

“After spending countless hours watching tutorial videos and reading more than 30 research papers about artificial intelligence and smart healthcare, I came up with a standard that can assist the health system,” he shared.

He said several vulnerabilities and adversarial attacks make it challenging for a safe and secure intelligent healthcare system from the security point of view.

“Machine learning has been used widely to develop suitable models to predict and mitigate attacks.

“Still, the attacks could trick the machine learning models and misclassify diseases and symptoms.

“As a result, it leads to incorrect decisions, for example, false disease detection and wrong treatment plans for patients,” said Arawinkumaar, who successfully completed his Master of IT majoring in Cyber Security and Networks at QUT in Brisbane last month.

His paper underlined the types of adversarial attacks and their impact on smart healthcare systems, and proposed a model to examine how such attacks impact machine learning classifiers.

To test the model that can classify medical images with high accuracy, Arawinkumaar used a medical image dataset.

“We then launched the model with a Fast Gradient Sign Method (FGSM) attack to cause the model to predict the images and misclassify them inaccurately.

“Using transfer learning, we trained a VGG-19 model with the medical dataset and later implement the FGSM to the Convolutional Neural Network (CNN) to examine the significant impact it causes on the performance and accuracy of the machine learning model.

“Our results demonstrate that the adversarial attack misclassifies the images, causing the model’s accuracy rate to drop from 88% to 11%,” said the former international student ambassador for the city of Brisbane.

Arawinkumaar’s work, which highlighted the defensive approaches and how hospitals can use intelligent healthcare systems to meet demands and solve health system challenges, was recently published at the 14th International Conference on Sensing Technology (ICST 2022) held at the Indian Institute of Technology Madras in Chennai from Jan 17 to 19.

His university had also invited him to publish his paper as a journal while he has his sights set on joining the IT workforce in Australia.