Topic 1: Metaverse

Motivation: The Metaverse, envisioned as a collective virtual shared space, is emerging at the confluence of various advanced technologies, including virtual and augmented reality, blockchain, and artificial intelligence. This digital realm promises an unprecedented level of integration between virtual environments and the physical world, offering new dimensions to social interactions, economic transactions, and digital identities. However, the ambitious development of the Metaverse brings forth significant technical challenges, such as creating interoperable and scalable virtual spaces, ensuring real-time interactions with minimal latency, and developing secure and transparent systems for digital assets and identities. These technical hurdles underscore the urgent need for focused research to address the complexities of building a seamless, accessible, and secure Metaverse. Such research is crucial not only for unlocking the full potential of these virtual spaces but also for anticipating and mitigating the social, ethical, and security implications that accompany the creation of such expansive digital universes.

Recent journal papers from our group on this topic:


Topic 2: Machine Learning for Cyberattack Detection

Motivation: Cyberattacks are a serious threat to the security and privacy of information systems and networks. They can cause significant damage to individuals, organizations, and society by compromising data, disrupting services, and stealing resources. To prevent and mitigate cyberattacks, various methods have been proposed to detect and classify them in real time. Among these methods, machine learning (ML) has emerged as a powerful and effective technique, as it can learn from data and adapt to changing environments. However, applying ML to cyberattack detection also poses many technical challenges, such as dealing with imbalanced and noisy data, selecting appropriate features and algorithms, ensuring robustness and scalability, and providing explainability and trustworthiness. These challenges require novel solutions and rigorous research from different disciplines and perspectives. Therefore, the research on machine learning for cyberattack detection is motivated by both the practical need to enhance cybersecurity and the scientific interest to advance ML theory and applications.

Recent journal papers  from our group on this topic:



Datasets: 

1) Blockchain Network Attack Traffic dataset (BNat): https://avitech-vnu.github.io/BNaT/#/


Topic 3: Machine Learning for Encrypted Data

Motivation: Machine learning (ML) is a powerful technique that can learn from data and provide insights, predictions, and recommendations for various domains and applications. However, ML often requires access to sensitive or confidential data, such as personal information, medical records, or financial transactions, which may pose privacy and security risks. How can we enable ML to operate on encrypted data without compromising its accuracy and efficiency? This is the main motivation for the research topic of machine learning for encrypted data, which explores novel cryptographic methods and algorithms that can perform ML tasks directly on ciphertexts, without revealing any information about the plaintexts. This research topic has significant implications for enhancing data protection and enabling privacy-preserving ML applications in various scenarios, such as cloud computing, federated learning, and edge computing.

Recent journal papers  from our group on this topic: