Modelling Long Range Dependencies in $N$D: From Task-Specific to a General Purpose CNN
This addresses the need for general-purpose CNNs that avoid task-specific tuning, offering a solution for researchers and practitioners working with diverse data types.
The paper tackles the problem of task-specific CNN architectures by introducing the Continuous Convolutional Neural Network (CCNN), a single model that processes data of arbitrary resolution, dimensionality, and length without structural changes, matching or outperforming state-of-the-art results across sequential, visual, and point-cloud tasks.
Performant Convolutional Neural Network (CNN) architectures must be tailored to specific tasks in order to consider the length, resolution, and dimensionality of the input data. In this work, we tackle the need for problem-specific CNN architectures. We present the Continuous Convolutional Neural Network (CCNN): a single CNN able to process data of arbitrary resolution, dimensionality and length without any structural changes. Its key component are its continuous convolutional kernels which model long-range dependencies at every layer, and thus remove the need of current CNN architectures for task-dependent downsampling and depths. We showcase the generality of our method by using the same architecture for tasks on sequential ($1{\rm D}$), visual ($2{\rm D}$) and point-cloud ($3{\rm D}$) data. Our CCNN matches and often outperforms the current state-of-the-art across all tasks considered.