Research Projects

Project 1: Paradigm Shift from Vague Legal Contracts to Blockchain-based Smart Contracts

As part of my broader research program, this project investigates the open methodological and implementation challenges involved in converting inherently vague legal contracts into accurate and enforceable smart legal contracts.

• Aim 1: Develop formal methodologies that consist of efficient models and algorithms that address current challenges in translating natural language contracts to smart legal contracts.
• Aim 2: Investigate and incorporate the Type-2 Fuzzy Logic and Decision Tree algorithms that help in decision-making for the selection of the most suitable interpretation from a set of ambiguous words and phrases.
• Aim 3: Explore on Contextual AI and NLP-based Fuzzy Logic and how it facilitates the automated generation of a realistic and feasible smart legal contract.
• Aim 4: Explore on how LLMs can be used to understand the vague legal contracts and rank the interpretations.

Project 2: Federated Learning for Secure and Trustworthy AI

This project focuses on strengthening the security and reliability of Federated Learning (FL), especially when training or deploying Large Language Models (LLMs). As FL becomes central to privacy-preserving AI, new risks such as malicious updates, unreliable nodes, and architectural weaknesses demand innovative solutions.

• Aim 1: Investigate security vulnerabilities in LLM-based federated learning and develop defenses to protect model updates and training integrity.
• Aim 2: Apply Game Theory to design incentive, reward, and penalty mechanisms that maintain honest participation and secure behavior among FL nodes.
• Aim 3: Develop a secure federated learning architecture with strong authentication, secure aggregation, and reliable model update verification.
• Aim 4: Explore the use of Blockchain or decentralized audit layers to create transparent, tamper-resistant logs of FL contributions and detect malicious activities across training rounds.

Project 3: Secure Metaverse Systems Using Multimodal AI and AR/VR

Metaverse is still in the germination phase, composed of various major technologies that work together and provide users with an unprecedented experience. Nevertheless, the Metaverse also has plenty of severe security threats and vulnerabilities that need immediate action. This project extends metaverse security research by developing real AR/VR applications while addressing emerging privacy, authenticity, and deepfake-driven threats in immersive environments. The goal is to combine multimodal deep learning with secure system design to support safe and trustworthy metaverse interactions.

• Aim 1: Design and develop AR/VR applications that integrate multimodal inputs (visual, audio, gesture, text) to create realistic, interactive, and context-aware metaverse experiences.
• Aim 2: Study security threats unique to immersive systems, such as avatar impersonation, sensor spoofing, spatial manipulation, and immersive phishing, and propose practical mitigation strategies.
• Aim 3: Explore multimodal deep learning techniques for detecting deepfakes, identity manipulation, and behavioral anomalies, and integrate blockchain for provenance tracking and secure auditing.
• Aim 4: Investigate how decentralized identity, smart contracts, and verifiable credentials can establish trust, authentication, and user control within AR/VR-based metaverse environments.

Project 4: Blockchain for Trusted Artificial Intelligence

As both Blockchain and Artificial Intelligence (AI) are two different popular technologies with two entirely different contributions, the goal of this project is to research how they both would complement each other, especially how Blockchain would benefit the integrity of the nodes, agents, models, and algorithms in artificial intelligence so that there is more trust and reliability in the decision-making process.

• Aim 1: Explore how the smart contract, gas, and transaction costs can be optimized when AI is integrated with Blockchain, as transactions performed in Blockchain are already expensive.
• Aim 2: Analyze the potential of developing and designing the model where all components are on-chain when AI is integrated with Blockchain to ensure its integrity. Additionally, the feasibility of this model would also be studied thoroughly, as reaching the objective of efficiency and optimization is essential.
• Aim 3: Explore the limitations of existing consensus mechanisms through which consensus among the nodes is achieved in Blockchain and develop new consensus mechanisms that overcome the challenges of existing consensus mechanisms, such as the requirement of only a 51% vote for approval of a decision.

Project 5: Data Cooperatives for Privacy and Security

Data Cooperatives are a novel area for secure data management which promises its users better protection and control of their personal data than the traditional way of their handling by the data collectors. Taking into account its advantages, I plan to research the development of the integration of Data Coop with Blockchain, challenges that are hindering its growth, and widespread use among data providers and consumers.

• Aim 1: Investigate the application of Blockchain on how data owners/providers can have better control of their personal data compared to the traditional approach where the big data companies handle their data for them.
• Aim 2: Explore the application of Game Theory in the design of the consensus mechanism as well as the reward and penalty system to maintain trust and authenticity in the data-sharing/validating agents.
• Aim 3: Investigate the security vulnerabilities that are associated with Data Cooperatives thoroughly and explore the mitigation strategies for each.