Imbedding Deep Neural Networks
This work addresses the theoretical and practical challenges of network depth in deep learning, offering a novel approach for researchers and practitioners, though it appears incremental in building upon Neural ODEs.
The paper tackles the problem of understanding and optimizing continuous-depth neural networks by introducing a new method based on Invariant Imbedding, which reduces the optimization to forward-facing initial value problems, showing competitive performance in supervised learning and time series prediction.
Continuous-depth neural networks, such as Neural ODEs, have refashioned the understanding of residual neural networks in terms of non-linear vector-valued optimal control problems. The common solution is to use the adjoint sensitivity method to replicate a forward-backward pass optimisation problem. We propose a new approach which explicates the network's `depth' as a fundamental variable, thus reducing the problem to a system of forward-facing initial value problems. This new method is based on the principle of `Invariant Imbedding' for which we prove a general solution, applicable to all non-linear, vector-valued optimal control problems with both running and terminal loss. Our new architectures provide a tangible tool for inspecting the theoretical--and to a great extent unexplained--properties of network depth. They also constitute a resource of discrete implementations of Neural ODEs comparable to classes of imbedded residual neural networks. Through a series of experiments, we show the competitive performance of the proposed architectures for supervised learning and time series prediction.