STEER : Simple Temporal Regularization For Neural ODE

Arnab Ghosh, HARKIRAT Behl, Emilien Dupont, Philip Torr, Vinay Namboodiri

Poster Session 7 (more posters)
on 2020-12-10T21:00:00-08:00 - 2020-12-10T23:00:00-08:00
GatherTown: Generative Models ( Town D1 - Spot D0 )
Join GatherTown
Only iff poster is crowded, join Zoom . Authors have to start the Zoom call from their Profile page / Presentation History.
Abstract: Training Neural Ordinary Differential Equations (ODEs) is often computationally expensive. Indeed, computing the forward pass of such models involves solving an ODE which can become arbitrarily complex during training. Recent works have shown that regularizing the dynamics of the ODE can partially alleviate this. In this paper we propose a new regularization technique: randomly sampling the end time of the ODE during training. The proposed regularization is simple to implement, has negligible overhead and is effective across a wide variety of tasks. Further, the technique is orthogonal to several other methods proposed to regularize the dynamics of ODEs and as such can be used in conjunction with them. We show through experiments on normalizing flows, time series models and image recognition that the proposed regularization can significantly decrease training time and even improve performance over baseline models.

Preview Video and Chat

Chat is not available.