Gary C. An

The paper introduces a novel agent-based model of the blood brain barrier to simulate its activation, injury, and restoration dynamics.

This paper models the innate immune responses of chickens and ducks to HPAIV, identifying RIG-I as a key molecular difference affecting immune outcomes.

This paper develops a digital twin platform, WEABM, to model VML healing and discover therapeutic strategies through simulation-based reinforcement learning.

The paper demonstrates generating synthetic cytokine time series data using a Critical Illness Digital Twin to improve ML/AI predictive performance for post-traumatic acute respiratory distress syndrome.

This paper uses simulation-based deep reinforcement learning to discover effective multimodal treatment strategies for systemic inflammation using existing immunomodulatory drugs.

This paper uses a convolutional Siamese neural network to predict wound functional states from images in a canine model, achieving MAPE between 5% and 50%.

This paper uses nested active learning to efficiently parameterize and contextualize an agent-based model of systemic inflammation, reducing required simulations by 99%.

This paper uses genetic algorithms to refine an agent-based model of inflammation to match clinical cytokine data from burn patients, resulting in a more biologically heterogeneous model.

Open-source agent-based PhysiCell model of SARS-CoV-2 infection in lung epithelium; version 5 clears virus in about 6 days in simulation, and delayed T cell arrival increases epithelial loss.

The paper presents a framework combining agent-based simulation and active learning to discover optimal cancer regression regions in immunotherapy models.

This paper introduces a high-throughput framework integrating PhysiCell and EMEWS to test cancer hypotheses through mechanistic simulations and optimize therapeutic strategies.

This paper uses a high-performance computing agent-based model of sepsis to identify novel metrics for characterizing the behavioral landscape of sepsis and highlights challenges in clinical data sparseness and reproducibility.

This paper uses genetic algorithms on an agent-based model of the immune system to explore multi-targeted control strategies for sepsis, showing that multi-point and time-dependent interventions are necessary.