Variational Inference of Parameters in Opinion Dynamics Models

2024-03-08

International Conference on Web and Social Media

AI Summary

Parameter estimation in agent-based models (ABMs) of opinion dynamics traditionally relies on costly simulation-based heuristics, hindering scalability and interpretability. Method: This paper introduces a differentiable probabilistic generative framework that systematically integrates variational inference into opinion dynamics ABMs for the first time. Leveraging Gumbel-Softmax reparameterization and normalizing flows as the variational family, it enables joint end-to-end optimization of macroscopic parameters (e.g., confidence thresholds) and high-dimensional microscopic role parameters (200-D). The inference task is formulated as a stochastic variational inference problem over a bounded-confidence model with role-aware agent heterogeneity. Contribution/Results: Our method significantly outperforms conventional simulation-based heuristics and MCMC approaches in accuracy, scalability, and computational efficiency. It achieves robust parameter recovery while preserving model interpretability—establishing a novel paradigm for explainable parameter learning in social simulation.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Probabilistic Circuits and Graphical Models

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Modeling human behavior through the lens of online social networks presents both a significant opportunity and a challenge for understanding complex social phenomena, such as misinformation spread, opinion formation and polarization. While agent-based models (ABMs) are widely used for studying these social phenomena, parameter estimation remains a challenge, often relying on costly simulation-based heuristics. This work uses variational inference to estimate the parameters of an opinion dynamics ABM by transforming the estimation problem into an optimization task that can be solved directly. Our proposal relies on probabilistic generative ABMs (PGABMs): we start by synthesizing a probabilistic generative model from the ABM rules. Then, we transform the inference process into an optimization problem suitable for automatic differentiation. In particular, we use the Gumbel-Softmax reparameterization for categorical agent attributes and Stochastic Variational Inference for parameter estimation. Moreover, we explore the trade-offs of using variational distributions with different complexities: Normal distributions and Normalizing Flows. We validate our method on a bounded confidence model with agent roles (leaders and followers), by estimating both macroscopic (bounded confidence intervals and backfire thresholds) and microscopic (200 categorical agent-level roles) parameters more accurately than simulation-based and MCMC methods.

Problem

• Estimates parameters in opinion dynamics agent-based models
• Transforms parameter inference into an optimization problem
• Enables data-driven validation of social behavior models

Innovation

• Variational inference for opinion dynamics ABM parameters
• Gumbel-Softmax reparameterization for categorical agent attributes
• Normalizing flows and normal distributions for variational approximations