The Unreasonable Effectiveness of Patches in Deep Convolutional Kernels Methods

Abstract

A recent line of work showed that various forms of convolutional kernel methods can be competitive with standard supervised deep convolutional networks on datasets like CIFAR-10, obtaining accuracies in the range of 87-90% while being more amenable to theoretical analysis. In this work, we highlight the importance of a data-dependent feature extraction step that is key to the obtain good performance in convolutional kernel methods. This step typically corresponds to a whitened dictionary of patches, and gives rise to a data-driven convolutional kernel methods. We extensively study its effect, demonstrating it is the key ingredient for high performance of these methods. Specifically, we show that one of the simplest instances of such kernel methods, based on a single layer of image patches followed by a linear classifier is already obtaining classification accuracies on CIFAR-10 in the same range as previous more sophisticated convolutional kernel methods. We scale this method to the challenging ImageNet dataset, showing such a simple approach can exceed all existing non-learned representation methods. This is a new baseline for object recognition without representation learning methods, that initiates the investigation of convolutional kernel models on ImageNet. We conduct experiments to analyze the dictionary that we used, our ablations showing they exhibit low-dimensional properties.

References in corpus (13)

• Do Deep Nets Really Need to be Deep?
• Neural Tangent Kernel: Convergence and Generalization in Neural Networks
• Do ImageNet Classifiers Generalize to ImageNet?
• A Downsampled Variant of ImageNet as an Alternative to the CIFAR datasets
• On Lazy Training in Differentiable Programming
• Convolutional Kernel Networks
• Scattering Networks for Hybrid Representation Learning
• Enhanced Convolutional Neural Tangent Kernels
• How to Scale Up Kernel Methods to Be As Good As Deep Neural Nets
• Neural Kernels Without Tangents
• Compressing the Input for CNNs with the First-Order Scattering Transform
• Greedy Layerwise Learning Can Scale to ImageNet
• On the Empirical Neural Tangent Kernel of Standard Finite-Width Convolutional Neural Network Architectures

Cited by in corpus (2)

• Convex Geometry and Duality of Over-parameterized Neural Networks
• Learning with convolution and pooling operations in kernel methods