Poster
in
Workshop: Methods and Opportunities at Small Scale (MOSS)
Neural Stochastic Differential Equations on Compact State-Spaces
Yue-Jane Liu · Malinda Lu · Matthew Nock · Yaniv Yacoby
Keywords: [ Generative Models ] [ Time Series ] [ Dynamical Systems ] [ Inductive Bias ] [ Viability Theory ] [ Neural SDEs ]
Many modern probabilistic models rely on SDEs, but their adoption is hampered by instability, poor inductive bias outside bounded domains, and reliance on restrictive dynamics or training tricks. While recent work constrains SDEs to compact spaces using reflected dynamics, these approaches lack continuous dynamics and efficient high-order solvers, limiting interpretability and applicability. We propose a novel class of neural SDEs on compact spaces with continuous dynamics, amenable to higher-order solvers and with favorable inductive bias.