Modern deep neural networks comprise many operational layers, such as dense or convolutional layers, which are often collected into blocks. In this work, we introduce a new, wavelet-transform-based network architecture that we call the multi-resolution perceptron: by adding a pooling layer, we create a new network block, the WavPool. The first step of the multi-resolution perceptron is transforming the data into its multi-resolution decomposition form by convolving the input data with filters of fixed coefficients but increasing size. Following image processing techniques, we are able to make scale and spatial information simultaneously accessible to the network without increasing the size of the data vector. WavPool outperforms a similar multilayer perceptron while using fewer parameters, and outperforms a comparable convolutional neural network by ~ 10% on relative accuracy on CIFAR-10.

翻译：现代深度神经网络由多个操作层组成，例如全连接层或卷积层，这些层通常被组织成模块。本文提出一种基于小波变换的新型网络架构，称为多分辨率感知器：通过引入池化层，我们创建了一个新的网络模块WavPool。多分辨率感知器的第一步是将输入数据与固定系数但尺寸递增的滤波器进行卷积，从而将数据转换为其多分辨率分解形式。借鉴图像处理技术，我们能够在无需增加数据向量尺寸的前提下，使网络同时获取尺度信息和空间信息。WavPool在参数更少的情况下性能优于同类多层感知器，且在CIFAR-10数据集上的相对准确率比可比较的卷积神经网络高出约10%。