Malliavin Calculus as Stochastic Backpropogation

We establish a rigorous connection between pathwise (reparameterization) and score-function (Malliavin) gradient estimators by showing that both arise from the Malliavin integration-by-parts identity. Building on this equivalence, we introduce a unified and variance-aware hybrid estimator that adaptively combines pathwise and Malliavin gradients using their empirical covariance structure. The resulting formulation provides a principled understanding of stochastic backpropagation and achieves minimum variance among all unbiased linear combinations, with closed-form finite-sample convergence bounds. We demonstrate 9% variance reduction on VAEs (CIFAR-10) and up to 35% on strongly-coupled synthetic problems. Exploratory policy gradient experiments reveal that non-stationary optimization landscapes present challenges for the hybrid approach, highlighting important directions for future work. Overall, this work positions Malliavin calculus as a conceptually unifying and practically interpretable framework for stochastic gradient estimation, clarifying when hybrid approaches provide tangible benefits and when they face inherent limitations.

💡 Research Summary

The paper establishes a rigorous theoretical bridge between the two dominant stochastic gradient estimators used in machine learning: the path‑wise (reparameterization) estimator and the score‑function (REINFORCE) estimator. By invoking the Malliavin integration‑by‑parts identity—specifically the duality between the Malliavin derivative (D) and the Skorokhod integral (\delta)—the authors prove that both estimators are special cases of the same underlying formula. For Gaussian distributions, they show that the path‑wise gradient (\nabla_\theta L(\theta)=\mathbb{E}_\varepsilon