About

Negin Alemazkoor is an Assistant Professor at the University of Virginia School of Engineering and Applied Science. Her research focuses on developing sensing and computing methodologies for the fast and reliable analysis of smart and interconnected infrastructure systems under uncertainty. Her work is inherently multi-disciplinary and aims to advance the reliability and resilience of infrastructure systems by facilitating fast and accurate system analysis, which leads to optimal system operation and management.

Research topics

• Computer Science
• Machine Learning
• Data Mining
• Economics
• Real-time computing
• Medicine
• Database
• Risk analysis (engineering)
• Environmental resource management
• Internal medicine
• Geography
• Ecology
• Business
• Meteorology
• Physics
• Intensive care medicine
• Oncology
• Engineering
• Environmental science
• Econometrics

Selected publications

• G-PARC: Graph-Physics Aware Recurrent Convolutional Neural Networks for Spatiotemporal Dynamics on Unstructured Meshes

  arXiv (Cornell University) · 2026-04-16

  preprintOpen access

  Physics-aware recurrent convolutional networks (PARC) have demonstrated strong performance in predicting nonlinear spatiotemporal dynamics by embedding differential operators directly into the computational graph of a neural network. However, pixel-based convolutions are restricted to static, uniform Cartesian grids, making them ill-suited to following evolving localized structures in an efficient manner. Graph neural networks (GNNs) naturally handle irregular spatial discretizations, but existing graph-based physics-aware deep learning (PADL) methods have difficulty handling extreme nonlinear regimes. To address these limitations, we propose Graph PARC (G-PARC), which uses moving least squares (MLS) kernels to approximate spatial derivatives on unstructured graphs, and embeds the derivatives of governing partial differential equations into the network's computational graph. G-PARC achieves better accuracy with 2-3x fewer parameters than MeshGraphNet, MeshGraphKAN, and GraphSAGE, replacing the traditional encoder-processor-decoder framework with analytically computed differential operators. We demonstrate that G-PARC (1) generalizes across nonuniform spatial and temporal discretizations; (2) handles moving meshes required for structural deformation; and (3) outperforms existing graph-based PADL methods on nonlinear benchmarks including fluvial hydrology, planar shock waves, and elastoplastic dynamics. By embedding explicit physical operators within the flexibility of GNNs, G-PARC enables accurate modeling of extreme nonlinear phenomena on complex computational domains, moving PADLbeyond idealized Cartesian grids.

  PublisherDOI

• Decision-Dependent Uncertainty-Aware Preventive-Corrective Strategy against Electrically Induced Wildfires

  SSRN Electronic Journal · 2026-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Decision-Dependent Uncertainty-Aware Preventive-Corrective Strategy against Electrically Induced Wildfires

  SSRN Electronic Journal · 2026-01-01

  preprintOpen accessSenior author
  PublisherDOI

• G-PARC: Graph-Physics Aware Recurrent Convolutional Neural Networks for Spatiotemporal Dynamics on Unstructured Meshes

  arXiv (Cornell University) · 2026-04-16

  articleOpen access

  Physics-aware recurrent convolutional networks (PARC) have demonstrated strong performance in predicting nonlinear spatiotemporal dynamics by embedding differential operators directly into the computational graph of a neural network. However, pixel-based convolutions are restricted to static, uniform Cartesian grids, making them ill-suited to following evolving localized structures in an efficient manner. Graph neural networks (GNNs) naturally handle irregular spatial discretizations, but existing graph-based physics-aware deep learning (PADL) methods have difficulty handling extreme nonlinear regimes. To address these limitations, we propose Graph PARC (G-PARC), which uses moving least squares (MLS) kernels to approximate spatial derivatives on unstructured graphs, and embeds the derivatives of governing partial differential equations into the network's computational graph. G-PARC achieves better accuracy with 2-3x fewer parameters than MeshGraphNet, MeshGraphKAN, and GraphSAGE, replacing the traditional encoder-processor-decoder framework with analytically computed differential operators. We demonstrate that G-PARC (1) generalizes across nonuniform spatial and temporal discretizations; (2) handles moving meshes required for structural deformation; and (3) outperforms existing graph-based PADL methods on nonlinear benchmarks including fluvial hydrology, planar shock waves, and elastoplastic dynamics. By embedding explicit physical operators within the flexibility of GNNs, G-PARC enables accurate modeling of extreme nonlinear phenomena on complex computational domains, moving PADLbeyond idealized Cartesian grids.

  PublisherOA PDF

• Multi-fidelity graph neural networks for efficient and accurate flood hazard mapping

  Environmental Modelling & Software · 2025-08-23 · 5 citations

  articleSenior authorCorresponding
  PublisherDOI

• Interpretable physics-informed graph neural networks for flood forecasting

  Computer-Aided Civil and Infrastructure Engineering · 2025-04-15 · 29 citations

  articleSenior author
  PublisherDOI

• End-to-End Graph Neural Networks for Real-Time Hydraulic Prediction in Stormwater Systems

  2025-09-01 · 1 citations

  articleOpen accessSenior authorCorresponding

  Abstract. Urban stormwater systems (SWS) play a critical role in protecting communities from pluvial flooding, ensuring public safety, and supporting resilient infrastructure planning. As climate variability intensifies and urbanization accelerates, there is a growing need for timely and accurate hydraulic predictions to support real-time control and flood mitigation strategies. While physics-based models such as SWMM provide detailed simulations of rainfall-runoff and flow routing processes, their computational demands often limit their feasibility for real-time applications. Surrogate models based on machine learning offer faster alternatives, but most rely on fully connected or grid-based architectures that struggle to capture the irregular spatial structure of drainage networks, often requiring precomputed runoff inputs and focusing only on node-level predictions. To address these limitations, we present GNN-SWS, a novel end-to-end graph neural network (GNN) surrogate model that emulates rainfall-driven hydraulic behavior across stormwater systems. The model predicts hydraulic states at both junctions and conduits directly from rainfall inputs, capturing the coupled dynamics of runoff generation and flow routing. It incorporates a spatiotemporal encoder–processor–decoder architecture with tailored message passing, autoregressive forecasting, and physics-guided constraints to improve predictive accuracy and physical consistency. Additionally, a training strategy based on the pushforward trick enhances model stability over extended prediction horizons. Applied to a real-world urban watershed, GNN-SWS demonstrates strong potential as a fast, scalable, and data-efficient alternative to traditional solvers. This framework supports key applications in urban flood risk assessment, real-time stormwater control, and the optimization of resilient infrastructure systems.

  PublisherOA PDFDOI

• Applications of Graph Neural Networks in Civil Infrastructures: A Review on Transportation, Power, Water, and Structural Systems

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Identifying key uncertainties in energy transitions with a Puerto Rico case study

  Nature Communications · 2025-10-13 · 2 citations

  articleOpen accessSenior author

  Deterministic energy transition planning risks uninformed decisions. Yet, the challenge of high-dimensional uncertainty-encompassing various technological, economic, social, and climatic factors-often leads to a deterministic treatment or simplification of uncertainties in planning. Here, we propose a computationally efficient framework that leverages surrogate-based sensitivity analysis to identify the key uncertainty sources driving the cost of different energy transition scenarios. We applied the proposed approach to Puerto Rico as a hurricane-prone power system that lacks efficient management. We find that changes in the frequency of hurricanes and organizational inefficiency are the two primary sources of uncertainty determining the system's total expected cost. When examining operational costs, different transition scenarios demonstrate unique key uncertainty sources. For example, the price of biofuel would mainly drive the operational cost when transitioning to a fully renewable power system. These findings can help planners by allowing them to focus on a narrower set of uncertainties in planning.

  PublisherDOI

• Deep learning-based downscaling of global digital elevation models for enhanced urban flood modeling

  Journal of Hydrology · 2025-01-16 · 25 citations

  article
  PublisherDOI

Frequent coauthors

• Hadi Meidani

  University of Illinois Urbana-Champaign

  12 shared

• Mazdak Tootkaboni

  10 shared

• Arghavan Louhghalam

  8 shared

• Mehdi Taghizadeh

  University of Virginia

  8 shared

• Mark Burris

  6 shared

• Harsh Anand

  6 shared

• Roshanak Nateghi

  Purdue University West Lafayette

  6 shared

• Md Abul Hasnat

  University of Virginia

  4 shared

Labs

• Alemzkoor Research GroupPI

Education

• PhD, Civil and Environmental Engineering

  University of Illinois at Urbana-Champaign

  2019