Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
129 tokens/sec
GPT-4o
28 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Inferring the origin of an epidemic with a dynamic message-passing algorithm (1303.5315v3)

Published 21 Mar 2013 in physics.soc-ph, cond-mat.stat-mech, cs.SI, and q-bio.PE

Abstract: We study the problem of estimating the origin of an epidemic outbreak -- given a contact network and a snapshot of epidemic spread at a certain time, determine the infection source. Finding the source is important in different contexts of computer or social networks. We assume that the epidemic spread follows the most commonly used susceptible-infected-recovered model. We introduce an inference algorithm based on dynamic message-passing equations, and we show that it leads to significant improvement of performance compared to existing approaches. Importantly, this algorithm remains efficient in the case where one knows the state of only a fraction of nodes.

Citations (272)

Summary

  • The paper presents a novel dynamic message-passing (DMP) algorithm that improves epidemic source estimation using probabilistic SIR modeling.
  • It outperforms traditional graph-centrality methods by efficiently handling incomplete network data across models like Erdős-Rényi and scale-free graphs.
  • The study’s linear-time approach offers practical advances for real-time epidemic control and paves the way for further dynamic network analysis.

Inferring the Origin of an Epidemic with a Dynamic Message-Passing Algorithm

The paper introduces a novel algorithm to estimate the origin of an epidemic outbreak within a network using the Susceptible-Infected-Recovered (SIR) model. This work is pivotal for understanding and controlling epidemic spreads across different types of networks, such as social or computer networks. The primary challenge tackled by this paper is accurately identifying the source of an epidemic, a task complicated by the stochastic nature of infection propagation.

Dynamic Message-Passing Algorithm

The core contribution of this paper is a dynamic message-passing (DMP) algorithm that offers a significant improvement over existing estimation methods. This algorithm calculates the probability that an observed outbreak originated from a specific node, even in cases where only a fraction of the network's state is observed. Unlike traditional methods that primarily use graph-centrality measures, this approach efficiently handles scenarios where infection information is incomplete.

The paper details how the DMP technique provides precise probabilistic results on tree-based structures, making it more reliable than conventional belief propagation (BP) algorithms. This is achieved by estimating marginal probabilities dynamically rather than relying on static assumptions about network states. The computational complexity of the DMP method is linear concerning the network size and time of simulation, making it a practical option for large-scale networks.

Performance and Evaluation

The algorithm's performance is assessed using various random network models, such as Erdős-Rényi and scale-free graphs, which are commonly used to simulate real-world network topologies. The results indicate that the DMP approach outperforms traditional estimators like distance and Jordan centrality in most tested scenarios, except in specific parameter ranges where deterministic models perform comparably well.

One of the key strengths of the DMP algorithm is its robustness in effectively estimating the epidemic source even when the snapshot of the network is incomplete. As a result, the DMP method retains its efficacy across different levels of observed network information, a feature not shared by centrality-based estimators.

Implications and Future Directions

The practical applications of this research are extensive. Accurate identification of epidemic origins can greatly enhance our ability to mitigate future outbreaks by implementing strategic containment and prevention measures. Theoretically, this work addresses a gap in dynamic network analysis by providing a probabilistic inference framework for time-evolving processes.

The paper leaves room for future explorations in extending the algorithm to handle networks with dynamic topology changes or to infer multiple sources of infection simultaneously. These extensions have tangible implications for improving real-time epidemic response strategies in interconnected systems.

In summary, the proposed dynamic message-passing algorithm opens new opportunities for precise epidemic source identification, showcasing substantial improvements over traditional methods. As networks continue evolving in complexity and scale, this approach provides a promising direction for enhancing our understanding of epidemic dynamics and designing effective countermeasures.