论文标题
插值技术加快神经源梯度的传播
Interpolation Technique to Speed Up Gradients Propagation in Neural ODEs
论文作者
论文摘要
我们提出了一种基于简单的基于插值的方法，用于在神经ode模型中有效地近似梯度。我们将其与反向动态方法（在文献中称为“伴随方法”）进行比较，以在分类，密度估计和推理近似任务上训练神经ODES。我们还提出了使用对数规范形式主义对我们的方法的理论理由。结果，我们的方法允许比反向动态方法更快的模型训练，该方法已通过多种标准基准的广泛数值实验确认和验证。
We propose a simple interpolation-based method for the efficient approximation of gradients in neural ODE models. We compare it with the reverse dynamic method (known in the literature as "adjoint method") to train neural ODEs on classification, density estimation, and inference approximation tasks. We also propose a theoretical justification of our approach using logarithmic norm formalism. As a result, our method allows faster model training than the reverse dynamic method that was confirmed and validated by extensive numerical experiments for several standard benchmarks.
