Towards Communication-Efficient Model Updating for On-Device Session-Based Recommendation

By X. Xia et al.

Published on June 10, 2023

Read the original document by opening this link in a new tab.

Table of Contents

1. Abstract
2. Introduction
3. Problem Setting
4. Methodology

Summary

The document discusses the challenges of updating on-device recommendation models efficiently to reduce network communication costs. It introduces a framework that combines model compression with update compression using compositional coding. Compositional coding represents items with discrete codes and learns embedding vectors for each code, allowing for greater modeling flexibility and expressive power. The proposed approach aims to maintain on-device models up-to-date with minimal communication costs.