AI and agentic AI for Healthcare

Healthcare is one of the places where machine learning needs to be useful, interpretable, and operationally realistic at the same time. My work in this topic studies health systems as networks and time-varying services: how patients move through referral pathways, how emergency departments become overcrowded, and how AI assistants can support triage and decision-making without hiding the reasoning process.

Across these projects, I combine graph neural networks, network science, time-series forecasting, causal explanations, and language-model based assistance. The common goal is to build models that can help clinicians and health-service managers understand bottlenecks, anticipate demand, and make safer decisions.

Publications in this topic

• C-PATH: Conversational Patient Assistance and Triage in Healthcare System, IEEE International Conference on Digital Health, 2025.
• Causal Inference Explanations for Graph Neural Networks, Causal@UAI2024, 2024.
• A Temporal Fusion Transformer for Long-Term Explainable Prediction of Emergency Department Overcrowding, ECML PKDD, 2023.
• Referral prediction in Healthcare using Graph Neural Networks, GReS at ACM RecSys, 2021.
• Dissecting medical referral mechanisms in health services using graph neural networks, Complex Networks, 2020.
• Understanding the Primary-Specialty Referral Mechanism using Network Science, Complex Networks, 2019.