{"id":"arxiv/2604.00333","url":"https://arxiv.org/abs/2604.00333","tags":["math.NA","cs.LG","physics.comp-ph"],"score":0,"title":"MVNN: A Measure-Valued Neural Network for Learning McKean-Vlasov Dynamics from Particle Data","domain":"math.NA","source":"arxiv","authors":["Liyao Lyu","Xinyue Yu","Hayden Schaeffer"],"reviews":0,"subject":"math.NA","summary":"Collective behaviors that emerge from interactions are fundamental to numerous biological systems. To learn such interacting forces from observations, we introduce a measure-valued neural network that infers measure-dependent interaction (drift) terms directly from particle-trajectory observations. The proposed architecture generalizes standard neural networks to operate on probability measures by learning cylindrical features, using an embedding network that produces scalable distribution-to-vector representations. On the theory side, we establish well-posedness of the resulting dynamics…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-04-01","reactions":{}}