Deep Function Machines: Generalized Neural Networks for Topological Layer Expression
This provides a foundational framework for designing resolution-invariant neural networks, which could benefit applications like high-resolution image processing.
The paper tackles the problem of neural networks being dependent on input dimension by proposing deep function machines (DFMs), a generalization that is invariant to input dimension, and introduces RippLeNet, which empirically achieves state-of-the-art invariance in computer vision.
In this paper we propose a generalization of deep neural networks called deep function machines (DFMs). DFMs act on vector spaces of arbitrary (possibly infinite) dimension and we show that a family of DFMs are invariant to the dimension of input data; that is, the parameterization of the model does not directly hinge on the quality of the input (eg. high resolution images). Using this generalization we provide a new theory of universal approximation of bounded non-linear operators between function spaces. We then suggest that DFMs provide an expressive framework for designing new neural network layer types with topological considerations in mind. Finally, we introduce a novel architecture, RippLeNet, for resolution invariant computer vision, which empirically achieves state of the art invariance.