Joint Learning of Network Topology and Opinion Dynamics Based on Bandit Algorithms
This addresses the challenge of modeling complex social interactions for researchers in network science and opinion dynamics, though it appears incremental as it builds on existing bandit algorithms for a specific domain.
The paper tackles the problem of jointly learning network topology and mixed opinion dynamics with diverse agent update rules, proposing a bandit-based algorithm that improves initial estimates, reduces prediction error, and outperforms methods like sparse linear regression and Gaussian process regression in numerical experiments.
We study joint learning of network topology and a mixed opinion dynamics, in which agents may have different update rules. Such a model captures the diversity of real individual interactions. We propose a learning algorithm based on multi-armed bandit algorithms to address the problem. The goal of the algorithm is to find each agent's update rule from several candidate rules and to learn the underlying network. At each iteration, the algorithm assumes that each agent has one of the updated rules and then modifies network estimates to reduce validation error. Numerical experiments show that the proposed algorithm improves initial estimates of the network and update rules, decreases prediction error, and performs better than other methods such as sparse linear regression and Gaussian process regression.