CKConv: Continuous Kernel Convolution For Sequential Data

Abstract

Conventional neural architectures for sequential data present important limitations. Recurrent networks suffer from exploding and vanishing gradients, small effective memory horizons, and must be trained sequentially. Convolutional networks are unable to handle sequences of unknown size and their memory horizon must be defined a priori. In this work, we show that all these problems can be solved by formulating convolutional kernels in CNNs as continuous functions. The resulting Continuous Kernel Convolution (CKConv) allows us to model arbitrarily long sequences in a parallel manner, within a single operation, and without relying on any form of recurrence. We show that Continuous Kernel Convolutional Networks (CKCNNs) obtain state-of-the-art results in multiple datasets, e.g., permuted MNIST, and, thanks to their continuous nature, are able to handle non-uniformly sampled datasets and irregularly-sampled data natively. CKCNNs match or perform better than neural ODEs designed for these purposes in a faster and simpler manner.

References in corpus (8)

• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• On the difficulty of training Recurrent Neural Networks
• Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains
• Searching for Activation Functions
• A Simple Way to Initialize Recurrent Networks of Rectified Linear Units
• Stand-Alone Self-Attention in Vision Models
• Complexity of Linear Regions in Deep Networks
• Generative Models as Distributions of Functions

Cited by in corpus (5)

• Efficiently Modeling Long Sequences with Structured State Spaces
• Combining Recurrent, Convolutional, and Continuous-time Models with Linear State-Space Layers
• FlexConv: Continuous Kernel Convolutions with Differentiable Kernel Sizes
• Generative Models as Distributions of Functions
• Exploiting Redundancy: Separable Group Convolutional Networks on Lie Groups