Universal Learning of Stochastic Dynamics for Exact Belief Propagation Using Bernstein Normalizing Flows
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (43) · 36610-36617
DOI: 10.1609/aaai.v40i43.40984 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Peter Amorese und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).
Abstract (Originalsprache)
Predicting the distribution of future states in a stochastic system, known as belief propagation, is fundamental to reasoning under uncertainty. However, nonlinear dynamics often make analytical belief propagation intractable, requiring approximate methods. When the system model is unknown and must be learned from data, a key question arises: can we learn a model that (i) universally approximates general nonlinear stochastic dynamics, and (ii) supports analytical belief propagation? This paper establishes the theoretical foundations for a class of models that satisfy both properties. The proposed approach combines the expressiveness of normalizing flows for density estimation with the analytical tractability of Bernstein polynomials. Empirical results show the efficacy of our learned model over state-of-the art data-driven methods for belief propagation, especially for highly non-linear systems with non-additive, non-Gaussian noise.
Zitieren
APA
Amorese, P., & Lahijanian, M. (2026). Universal Learning of Stochastic Dynamics for Exact Belief Propagation Using Bernstein Normalizing Flows. Proceedings of the AAAI Conference on Artificial Intelligence, 40(43), 36610-36617. https://doi.org/10.1609/aaai.v40i43.40984
Vancouver
Amorese P, Lahijanian M. Universal Learning of Stochastic Dynamics for Exact Belief Propagation Using Bernstein Normalizing Flows. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(43):36610-36617. doi:10.1609/aaai.v40i43.40984
BibTeX
@article{Amorese2026,
author = {Amorese, Peter and Lahijanian, Morteza},
title = {Universal Learning of Stochastic Dynamics for Exact Belief Propagation Using Bernstein Normalizing Flows},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {43},
pages = {36610--36617},
doi = {10.1609/aaai.v40i43.40984}
}
EUVIMED