A Fully Trainable Network with RNN-based Pooling

TL;DR

This paper introduces a fully trainable neural network with an RNN-based pooling layer that adapts to data, improving performance especially in small networks, demonstrated on CIFAR-10.

Contribution

It proposes a novel learnable pooling function using RNNs, enabling the pooling component to be fully trained and data-adaptive, unlike traditional handcrafted pooling methods.

Findings

RNN-based pooling approximates traditional pooling functions effectively

FTN improves small network performance by 7% error rate on CIFAR-10

Learnable pooling enhances overall network adaptability and accuracy

Abstract

Pooling is an important component in convolutional neural networks (CNNs) for aggregating features and reducing computational burden. Compared with other components such as convolutional layers and fully connected layers which are completely learned from data, the pooling component is still handcrafted such as max pooling and average pooling. This paper proposes a learnable pooling function using recurrent neural networks (RNN) so that the pooling can be fully adapted to data and other components of the network, leading to an improved performance. Such a network with learnable pooling function is referred to as a fully trainable network (FTN). Experimental results have demonstrated that the proposed RNN-based pooling can well approximate the existing pooling functions and improve the performance of the network. Especially for small networks, the proposed FTN can improve the performance by seven percentage points in terms of error rate on the CIFAR-10 dataset compared with the traditional CNN.