Neural SDE: Stabilizing Neural ODE Networks with Stochastic Noise

Neural Ordinary Differential Equation (Neural ODE) has been proposed as a continuous approximation to the ResNet architecture. Some commonly used regularization mechanisms in discrete neural networks (e.g. dropout, Gaussian noise) are missing in current Neural ODE networks. In this paper, we propose a new continuous neural network framework called Neural Stochastic Differential Equation (Neural SDE) network, which naturally incorporates various commonly used regularization mechanisms based on random noise injection. Our framework can model various types of noise injection frequently used in discrete networks for regularization purpose, such as dropout and additive/multiplicative noise in each block. We provide theoretical analysis explaining the improved robustness of Neural SDE models against input perturbations/adversarial attacks. Furthermore, we demonstrate that the Neural SDE network can achieve better generalization than the Neural ODE and is more resistant to adversarial and non-adversarial input perturbations.

Neural SDE: Stabilizing Neural ODE Networks with Stochastic Noise Xuanqing Liu Tesi Xiao Department of Computer Science, UCLA Department of Statistics, UC Davis xqliu@cs.ucla.edu texiao@ucdavis.edu arXiv:1906.02355v1 [cs.LG] 5 Jun 2019…

related reading

• Scalable Gradients for Stochastic Differential Equationsarxiv.org
• SDE-Net: Equipping Deep Neural Networks with Uncertainty Estimatesarxiv.org
• Neural Ordinary Differential Equationsarxiv.org
• [2604.06366] Stochastic Gradient Descent in the Saddle-to-Saddle Regime of Deep Linear Networksarxiv.org
• Neural Stochastic Differential Equations: Deep Latent Gaussian Models in the Diffusion Limitarxiv.org
• Stochastic Differential Equations, Deep Learning, and High-Dimensional PDEsmitmath.github.io
• 1806.07366arxiv.org
• [2605.01172] A Theory of Generalization in Deep Learningarxiv.org
• Neural Manifold Ordinary Differential Equationsarxiv.org
• Stochastic Parameter Decompositionarxiv.org
• An SDE Framework for Adversarial Training, with Convergence and Robustness Analysisarxiv.org
• Denoising Density Estimatorsarxiv.org