Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DEpiABS: Differentiable Epidemic Agent-Based Simulator

Zhijian Gao, Shuxin Li, Bo An

TL;DR

This work tackles the triad of fidelity, efficiency, and interpretability in epidemic modelling by introducing DEpiABS, a structure-centric differentiable ABM that captures fine-grained dynamics including resource constraints, mutation, and reinfection within a fully transparent framework. It couples end-to-end differentiability with a z-score-based scaling method to calibrate micro-level mechanisms to real data and to project outputs to large populations without losing granularity. Through sensitivity analyses and cross-regional calibration, DEpiABS demonstrates improved forecasting accuracy and data efficiency relative to state-of-the-art data-centric DABMs, while maintaining mechanistic interpretability. The approach achieves linear scalability with population size and forecast horizon, offering a practical and generalisable tool for future epidemic response planning.

Abstract

The COVID-19 pandemic highlighted the limitations of existing epidemic simulation tools. These tools provide information that guides non-pharmaceutical interventions (NPIs), yet many struggle to capture complex dynamics while remaining computationally practical and interpretable. We introduce DEpiABS, a scalable, differentiable agent-based model (DABM) that balances mechanistic detail, computational efficiency and interpretability. DEpiABS captures individual-level heterogeneity in health status, behaviour, and resource constraints, while also modelling epidemic processes like viral mutation and reinfection dynamics. The model is fully differentiable, enabling fast simulation and gradient-based parameter calibration. Building on this foundation, we introduce a z-score-based scaling method that maps small-scale simulations to any real-world population sizes with negligible loss in output granularity, reducing the computational burden when modelling large populations. We validate DEpiABS through sensitivity analysis and calibration to COVID-19 and flu data from ten regions of varying scales. Compared to the baseline, DEpiABS is more detailed, fully interpretable, and has reduced the average normal deviation in forecasting from 0.97 to 0.92 on COVID-19 mortality data and from 0.41 to 0.32 on influenza-like-illness data. Critically, these improvements are achieved without relying on auxiliary data, making DEpiABS a reliable, generalisable, and data-efficient framework for future epidemic response modelling.

DEpiABS: Differentiable Epidemic Agent-Based Simulator

TL;DR

This work tackles the triad of fidelity, efficiency, and interpretability in epidemic modelling by introducing DEpiABS, a structure-centric differentiable ABM that captures fine-grained dynamics including resource constraints, mutation, and reinfection within a fully transparent framework. It couples end-to-end differentiability with a z-score-based scaling method to calibrate micro-level mechanisms to real data and to project outputs to large populations without losing granularity. Through sensitivity analyses and cross-regional calibration, DEpiABS demonstrates improved forecasting accuracy and data efficiency relative to state-of-the-art data-centric DABMs, while maintaining mechanistic interpretability. The approach achieves linear scalability with population size and forecast horizon, offering a practical and generalisable tool for future epidemic response planning.

Abstract

The COVID-19 pandemic highlighted the limitations of existing epidemic simulation tools. These tools provide information that guides non-pharmaceutical interventions (NPIs), yet many struggle to capture complex dynamics while remaining computationally practical and interpretable. We introduce DEpiABS, a scalable, differentiable agent-based model (DABM) that balances mechanistic detail, computational efficiency and interpretability. DEpiABS captures individual-level heterogeneity in health status, behaviour, and resource constraints, while also modelling epidemic processes like viral mutation and reinfection dynamics. The model is fully differentiable, enabling fast simulation and gradient-based parameter calibration. Building on this foundation, we introduce a z-score-based scaling method that maps small-scale simulations to any real-world population sizes with negligible loss in output granularity, reducing the computational burden when modelling large populations. We validate DEpiABS through sensitivity analysis and calibration to COVID-19 and flu data from ten regions of varying scales. Compared to the baseline, DEpiABS is more detailed, fully interpretable, and has reduced the average normal deviation in forecasting from 0.97 to 0.92 on COVID-19 mortality data and from 0.41 to 0.32 on influenza-like-illness data. Critically, these improvements are achieved without relying on auxiliary data, making DEpiABS a reliable, generalisable, and data-efficient framework for future epidemic response modelling.
Paper Structure (33 sections, 15 equations, 9 figures, 12 tables, 2 algorithms)

This paper contains 33 sections, 15 equations, 9 figures, 12 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Traditional models
  4. Agent-based models
  5. Differentiable agent-based models
  6. Problem Description
  7. Model Structure
  8. Society
  9. Epidemic
  10. Population
  11. Model Workflow
  12. Parameter Calibration Method
  13. Differentiable Approximation
  14. Relaxation
  15. Tensorisation
  16. ...and 18 more sections

Figures (9)

  • Figure 1: Health class transition
  • Figure 2: Model Workflow
  • Figure 3: Cumulative infections under different transmission probabilities (left), mean monthly bill (middle), and maximum purchasable daily quotas of supplies (right)
  • Figure 4: Cumulative mortality under different means (left) and standard deviations (right) of symptom progression
  • Figure 5: Cumulative agents in critical conditions under different means (left) and standard deviations (right) of symptom progression
  • ...and 4 more figures