Improving the Accuracy of Early Exits in Multi-Exit Architectures via Curriculum Learning

TL;DR

This paper introduces Multi-Exit Curriculum Learning, a training strategy that improves the accuracy of early exits in multi-exit neural network architectures, especially useful for resource-constrained, time-sensitive applications.

Contribution

The paper proposes a novel curriculum learning method tailored for multi-exit architectures to enhance early exit accuracy during inference.

Findings

Consistent accuracy improvements on CIFAR-10 and CIFAR-100 datasets.

Effective across various multi-exit architecture configurations.

Abstract

Deploying deep learning services for time-sensitive and resource-constrained settings such as IoT using edge computing systems is a challenging task that requires dynamic adjustment of inference time. Multi-exit architectures allow deep neural networks to terminate their execution early in order to adhere to tight deadlines at the cost of accuracy. To mitigate this cost, in this paper we introduce a novel method called Multi-Exit Curriculum Learning that utilizes curriculum learning, a training strategy for neural networks that imitates human learning by sorting the training samples based on their difficulty and gradually introducing them to the network. Experiments on CIFAR-10 and CIFAR-100 datasets and various configurations of multi-exit architectures show that our method consistently improves the accuracy of early exits compared to the standard training approach.