Continual Learning at the Edge: Real-Time Training on Smartphone Devices

TL;DR

This paper presents a hybrid continual learning strategy, AR1*, enabling real-time personalized model training on smartphones without forgetting, addressing efficiency and data distribution challenges.

Contribution

Introduction of AR1*, a hybrid continual learning method, for on-device personalization on smartphones, demonstrating effective real-time training without catastrophic forgetting.

Findings

AR1* achieves effective personalization on Android devices.

The approach prevents catastrophic forgetting during on-device training.

Benchmark results show improved efficiency and effectiveness.

Abstract

On-device training for personalized learning is a challenging research problem. Being able to quickly adapt deep prediction models at the edge is necessary to better suit personal user needs. However, adaptation on the edge poses some questions on both the efficiency and sustainability of the learning process and on the ability to work under shifting data distributions. Indeed, naively fine-tuning a prediction model only on the newly available data results in catastrophic forgetting, a sudden erasure of previously acquired knowledge. In this paper, we detail the implementation and deployment of a hybrid continual learning strategy (AR1*) on a native Android application for real-time on-device personalization without forgetting. Our benchmark, based on an extension of the CORe50 dataset, shows the efficiency and effectiveness of our solution.