Research Symposium Highlights Innovations in Cybersecurity and AI

In an electrifying display of cutting-edge research, KFSCIS’ Research Symposium left attendees in awe as groundbreaking studies took center stage. Among the myriad of presentations, several posters and presentations stood out, pushing the boundaries of innovative research. Graduate students presented posters and conducted “lightning rounds” highlighting their most recent scientific research. The Symposium showcased innovations in cybersecurity, AI, data science, bioinformatics and health, robotics and underwater vehicles, federated learning, and climate research.

Three KFSCIS Ph.D. students received top honors at the symposium for their work in cybersecurity and AI. The students, Maryna Veksler, Harun Oz, and Mahshad Shariatnasab are Ph.D. students working under the mentorship of Dr. Kemal Akkaya and Dr. Selcuk Uluagac, and Dr. Farhad Shirani respectively.

Maryna Vekslar presented her work on threat analysis of Generative Adversarial Networks (GAN)-generated drone sensor data against state-of-the-art ML-based threat anomaly detectors. Her poster showcased their groundbreaking work, focusing on a novel GAN architecture designed to generate fake sensor data for accelerometer and gyroscope IMU sensors. In her work, she proposed developing an ensemble of GANs to capture spatial and temporal relationships between the sensors. Scientific results indicate that using an ensemble of GANs allows for a significant increase in the attack success rate against autoencoder and Convolutional Neural Network (CNN) anomaly detectors compared to when the data is generated in isolation using separate GANs. She will continue her work in researching and developing more sophisticated targeted attacks and analyzing security mechanisms to improve sensor anomaly detection.

Harun Oz introduced research on a newly discovered attack vector, which could potentially increase the risk of ransomware attacks. For the first time in the literature, they extensively studied a browser-based ransomware called RøB. Utilizing the FSA API and WebAssembly technologies, the researchers demonstrated RøB’s ability to encrypt user files directly from the browser, marking a paradigm shift in ransomware’s malicious tactics. The team’s analysis demonstrates that RøB can encrypt the victim’s local files including cloud-integrated directories, external storage devices, and network-shared folders regardless of the access limitations imposed by the API. Existing defense solutions fall short against this new threat. Harun and his team proposed three potential defense solutions to mitigate this new attack vector, which has recently been published in a conference paper at the 32nd USENIX Security Symposium.

The students were recognized for their research work, receiving certificates of recognition at the KFSCIS Annual Awards Celebration.

// Live helper chat widget begin // Live helper chat widget end