Abstract:We study the tipping point collective dynamics of an adaptive susceptible-infected-susceptible (SIS) epidemiological network in a data-driven, machine learning-assisted manner. We identify a parameter-dependent effective stochastic differential equation (eSDE) in terms of physically meaningful coarse mean-field variables through a deep-learning ResNet architecture inspired by numerical stochastic integrators. We construct an approximate effective bifurcation diagram based on the identified drift term of the eSDE and contrast it with the mean-field SIS model bifurcation diagram. We observe a subcritical Hopf bifurcation in the evolving network's effective SIS dynamics, that causes the tipping point behavior; this takes the form of large amplitude collective oscillations that spontaneously -- yet rarely -- arise from the neighborhood of a (noisy) stationary state. We study the statistics of these rare events both through repeated brute force simulations and by using established mathematical/computational tools exploiting the right-hand-side of the identified SDE. We demonstrate that such a collective SDE can also be identified (and the rare events computations also performed) in terms of data-driven coarse observables, obtained here via manifold learning techniques, in particular Diffusion Maps. The workflow of our study is straightforwardly applicable to other complex dynamics problems exhibiting tipping point dynamics.

What problem does this paper attempt to address?

The paper attempts to address the problem of identifying critical point dynamics in evolutionary epidemiological networks and modeling this process through a data-driven approach assisted by machine learning. Specifically, the researchers focus on the critical point dynamics in an adaptive Susceptible-Infected-Susceptible (SIS) epidemiological network model. This network model considers social interactions between individuals and allows susceptible individuals to avoid contact with infected individuals by reconnecting edges connected to infected individuals. The researchers use a data-driven approach, employing deep learning algorithms (such as those based on the ResNet architecture) to identify a parameter-dependent effective stochastic differential equation (eSDE) and construct an approximate effective bifurcation diagram through the drift term of this equation. The study finds that there is a subcritical Hopf bifurcation point in the effective SIS dynamics of the evolutionary network, leading to critical point behavior characterized by spontaneous but rare large-scale collective oscillations near the noise-stabilized state. Furthermore, the researchers demonstrate how to use diffusion map techniques to extract coarse-grained observable variables from the data and identify an eSDE based on these variables. By analyzing the eSDE, the researchers can perform computer-assisted rare event analysis (escape time computation) of the "catastrophic jumps" from the stable state to large-scale collective oscillations and compare it with the full epidemiological network model. The results show that using the eSDE model can significantly reduce the computational effort required to calculate escape times. In summary, this paper aims to identify and understand the critical point dynamics in evolutionary epidemiological networks through a data-driven approach and proposes effective modeling and analysis tools.

Tipping Points of Evolving Epidemiological Networks: Machine Learning-Assisted, Data-Driven Effective Modeling

Tipping points of evolving epidemiological networks: Machine learning-assisted, data-driven effective modeling

Epidemic Dynamics on Information-Driven Adaptive Networks

On the Dynamics of Deterministic Epidemic Propagation over Networks

Mean-Field Modeling Approach for Understanding Epidemic Dynamics in Interconnected Networks

Efficient weighted-ensemble network simulations of the SIS model of epidemics

Dynamical Processes and Epidemic Threshold on Nonlinear Coupled Multiplex Networks

Epidemionics: from the host-host interactions to the systematic analysis of the emergent macroscopic dynamics of epidemic networks

Dynamics of Simplicial SEIRS Epidemic Model: Global Asymptotic Stability and Neural Lyapunov Functions.

Using machine learning to identify epidemic threshold in complex networks

Coupled dynamics of endemic disease transmission and gradual awareness diffusion in multiplex networks

Dynamic Behavior of the Interaction Between Epidemics and Cascades on Heterogeneous Networks

Simulated identification of epidemic threshold on finite-size networks

SIR Epidemics in Interconnected Networks: threshold curve and phase transition

Epidemic Outbreaks on Networks with Effective Contacts

Edge-based Epidemic Dynamics with Multiple Routes of Transmission on Random Networks

Tipping point and noise-induced transients in ecological networks

Epidemics on networks with heterogeneous population and stochastic infection rates

Task-oriented machine learning surrogates for tipping points of agent-based models

An overview of epidemic models with phase transitions to absorbing states running on top of complex networks

An Exact Relationship Between Invasion Probability and Endemic Prevalence for Markovian SIS Dynamics on Networks