External Research Grants

This project proposes the development of GeneralGAD, a next-generation system for detecting anomalies in complex networks such as financial transactions, social interactions, and molecule networks. Unlike existing approaches that are often domain-specific and heavily reliant on large labeled datasets, GeneralGAD aims to operate as a general-purpose solution capable of identifying unusual patterns across diverse settings with minimal or no task-specific training data.

Technically, the project will integrate graph neural networks, which capture structural relationships in data, with large language models that provide contextual reasoning capabilities. This combination will enable the system to learn transferable representations of normal and abnormal behaviors across heterogeneous graphs and to generalize effectively to previously unseen domains. Building on promising preliminary results, the project will scale these models to enhance robustness, adaptability, and interpretability.

The potential impact is broad and tangible. GeneralGAD could transform how organizations detect fraud, mitigate cyber threats, and monitor harmful online activities, while also supporting discovery in scientific and medical research. By reducing dependence on costly labeled data and frequent retraining, the system offers a more efficient and scalable approach to safeguarding digital and real-world systems, ultimately contributing to greater security, trust, and innovation across industries.

The research aims to uncover and empirically test the overlooked psychological mechanisms that sustain age discrimination in Singapore, with the goal of informing evidence-based interventions. Specifically, the project will (1) Investigate the psychological impact of aging-related accommodations; (2) Develop precommitment strategies to increase the adoption of assistive technology; and (3) Determine how cultural tightness constrains older consumers.

This proposal aims to propose, design, and evaluate the economic and environmental performance of a business model for vehicle-to grid (V2G), where electricity flows from electric vehicles (EVs) to the electric grid. In this business model, each EV driver performs V2G services a few times (e.g., five) a year during peak electricity demand, which ensures negligible battery degradation, alleviating concerns over the battery life expectancy for both drivers and manufacturers. We will study EV drivers’ behaviors via a V2G pilot program and assess and optimize the performance of our business model with data from the pilot program.

The Singapore International Dispute Resolution Academy (SIDRA) is a platform for thought leadership in international dispute resolution theory, practice and policy, and will conduct the following research programmes for the next few years:
•    Dispute Resolution Empirical Research
•    International Mediation and the Singapore Convention on Mediation
•    Next Generation Dispute Resolution

PerFormRect aims to detect the misconceptions of students and utilize these insights in a feedback loop for crafting personalized formative assessment questions in programming education using Large Language Models (LLM). This project addresses Theme 4 (Leveraging Technology to Enhance and Personalize the Learning Experience) with efficient analysis of individual coding misconceptions and timely intervention through personalized coding questions for both practice and self-assessment. It tackles the pressing need to efficiently identify coding misconceptions of students and provide timely, personalized questions for both practice and self-assessment by students – tasks which are difficult for human instructors to perform at scale. PerformRect leverages on LLM to identify students’ misconceptions and generate personalized feedback (WP1) based on their code submissions. The identified misconceptions will also be used to generate formative assessment coding questions (WP2) tailored for their repetitive learning. The effectiveness of PerFormRect on students’ learning will be evaluated using randomized control trials (RCTs) in two institutions (for generalizability). In all, PerFormRect offers on-demand access, personalized feedback, tailored assessments and scalable programming skills development opportunities, allowing for a broader reach without sacrificing quality.

This project empowers educators to effectively leverage Generative AI as a transformative tool that not only identifies and addresses student misconceptions but also deepens understanding, fosters critical thinking, and enriches the overall learning experience. It proposes a novel AI-enhanced pedagogical approach called Debunkr, which utilises a cognitive conflict instructional approach to actively debunk misconceptions for university courses, potentially adaptable for any subject in the age of Generative AI.

TITAN 2.0 is a 2-year project to build an advanced AI-driven framework that identifies and fixes complex security flaws in software code. By combining traditional static analysis with the reasoning power of large language models (LLMs) orchestrated in an agentic fashion, the system extends analysis beyond individual functions to comprehend complex interactions across multiple files and modules. This allows it to catch interprocedural vulnerabilities that simpler LLM-powered tools often miss. Designed to support multiple programming languages such as Java, C#, Python, and JavaScript, the framework does not just flag risks; it acts as a digital security partner by providing automated CWE labeling, validated code patches, and developer-friendly reports. By integrating these smart agents, the project significantly improves vulnerability remediation to ensure that digital services remain secure and resilient.

This research / project is supported by the National Research Foundation, Singapore, and Ministry of Digital Development & Information under its Smart Nation and Digital Government Translational R&D Grant (Award No: TRANS2026-TGC01).

The project examines what Southeast Asian regional resilience consists of with respect to six contemporary challenges: creating regional order, maintaining supply chains, instituting soft law, managing digital threats, dealing with health pandemics, and managing demographic transitions. By discovering the ingredients that constitute regional resilience in these areas, the project seeks to provide insights on how Singapore and ASEAN can enhance their ability to withstand and recover from external shocks and perturbations to the region. This project is supported by the Social Science Research Council and will be done in collaboration with faculty members from the National University of Singapore (NUS) and the Nanyang Technological University (NTU).

This project explores what inclusive AI governance is, or could be, in six of Southeast Asia’s largest cities. It considers how Artificial Intelligence (AI) is reshaping urban governance, and how urban governance is shaping AI deployments. It argues that inclusion is not a technical feature of AI systems but a relational outcome of governance processes. It advances an Inclusive AI Governance Framework that links upstream policy design and regulation with downstream implementation and lived experience, analysing how institutional capacity, regulatory priorities, and local socio-economic conditions shape whether AI might mitigate or exacerbate urban inequalities. Through comparative policy mapping, ethnographic research, and cross-sectoral case studies in education, healthcare, mobility, and urban planning, the project will generate both theoretical insights and applied tools – including a framework, taxonomy, and handbook – to help policymakers and practitioners assess governance readiness, manage inclusion risks, and adapt global AI principles to diverse regional contexts.

This project is supported by the Social Science Research Council and will be done in collaboration with faculty members from the National University of Singapore (NUS) and the Nanyang Technological University (NTU).

This project is a complete end-to-end study on a more targeted approach to issue lightning risk alerts for ground personnel at Changi Airport to enable more granular decision-making. The project will include an economic study of operational disruption at the airport due to a lightning warning and event.