Piecewise-constant Neural ODEs
This incremental improvement addresses efficiency issues for researchers and practitioners using Neural ODEs in sequential data modeling, particularly in synthetic physics and planning tasks.
The paper tackles the computational challenges of Neural ODEs by proposing a piecewise-constant approximation, which enables exact Euler integration and reduces autoregressive sampling steps by 3-20 times compared to RNN and ODE-RNN models while matching baseline performance with less training and evaluation time.
Neural networks are a popular tool for modeling sequential data but they generally do not treat time as a continuous variable. Neural ODEs represent an important exception: they parameterize the time derivative of a hidden state with a neural network and then integrate over arbitrary amounts of time. But these parameterizations, which have arbitrary curvature, can be hard to integrate and thus train and evaluate. In this paper, we propose making a piecewise-constant approximation to Neural ODEs to mitigate these issues. Our model can be integrated exactly via Euler integration and can generate autoregressive samples in 3-20 times fewer steps than comparable RNN and ODE-RNN models. We evaluate our model on several synthetic physics tasks and a planning task inspired by the game of billiards. We find that it matches the performance of baseline approaches while requiring less time to train and evaluate.