mitmedialab / DeepABM-Pandemic

DeepABM: Scalable and Differentiable Agent-based Modeling of Epidemiologic Dynamics and Interventions

Implementation Reference

https://arxiv.org/pdf/2110.04421.pdf

Performance benchmarking

• 100,000 agents and 2 million interactions per second (to include graphic!)

Interventions Supported

• Clinical: Two-dose Vaccination (delay days, efficacy), Testing (specificity, speed)
• Digital: Contact Tracing (adoption rate, compliance probability)
• Behavioral - Quarantine (days, break probability)

Code used in the following papers:

1. Ayush Chopra et al: DeepABM - Scalable and Efficient Agent-based Simulations. Winter Simulation Conference 2021
2. Romero-brufau, Ayush Chopra et al: Public Health Impact of Delaying 2nd dose of COVID-mRNA vaccine. British Medical Journal 2021.
3. Gauri Gupta, Ritvik Kapila, Ayush Chopra, Ramesh Raskar: First 100 days of a pandemic - an interplay of clinical, behavioral and digital interventions. AAMAS 2024

Citation

@article{chopra2021deepabm, title={DeepABM: scalable, efficient and differentiable agent-based simulations via graph neural networks}, author={Chopra, Ayush and Gel, Esma and Subramanian, Jayakumar and Krishnamurthy, Balaji and Romero-Brufau, Santiago and Pasupathy, Kalyan S and Kingsley, Thomas C and Raskar, Ramesh}, journal={arXiv preprint arXiv:2110.04421}, year={2021} }

Contact

Please email: [ayush + c]@[mit.edu]