{"id":"arxiv/2610.17369","url":"https://arxiv.org/abs/2610.17369","tags":["math.PR"],"score":0,"title":"Large Network limit for Latent Variable Process Models with Endogeneously Evolving Interactions among Multiple Types","domain":"math.PR","source":"arxiv","authors":["Ankan Ganguly","Konstantinos Spiliopoulos"],"reviews":0,"subject":"math.PR","summary":"Motivated by the opinion and social dynamics literature, we rigorously study dynamically co-evolving latent space network models, where agent opinions evolve based on a latent state space model influenced by and influencing an underlying dynamically evolving graph determining their interactions. In contrast to earlier work, see [18], agents are allowed to be heterogeneous, i.e., of different types, and their dynamic evolution can capture behaviors like attraction and repulsion between and within groups. In addition, the model encodes bidirectional feedback between the latent process and the…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-12","reactions":{}}