This area studies graph neural networks, relational reasoning, explainable graph learning, and network-aware recommendation. The work emphasizes structural representation, generalization, and interpretable decision making.

This stream studies AI-enabled surveillance of market abuse, shadow trading, and cross-product manipulation. The work combines graph learning, spatio-temporal modeling, and regulatory analytics for modern financial markets.

These projects apply visual learning, representation learning, and world models to infrastructure inspection, energy systems, and Earth observation. The work targets robust perception under limited labels and challenging deployment conditions.

This stream investigates fairness, bias mitigation, and trustworthy AI for socially sensitive domains. The goal is to make predictive systems more transparent, equitable, and reliable in practice.

This category brings together forecasting, optimization, and resilience analysis for power, water, and utility systems. The work emphasizes operational decision support, carbon-aware planning, and resilient service delivery.

This category covers sequential decision processes, temporal point processes, and interpretable event modeling. It supports better understanding of dynamic systems, workflows, and complex interaction patterns.