Efficient Graph based Recommender System with Weighted Averaging of Messages

TL;DR

This paper introduces a scalable graph-based recommendation system that addresses perpetual cold start issues by reducing dataset size and employing a novel weighted message averaging technique, significantly improving efficiency and recall.

Contribution

The authors propose WAML, a new message aggregation method that reduces computational costs and enhances recommendation accuracy in large, cold-started product graphs.

Findings

WAML reduces training compute time to 1/7 of LightGCN.

WAML increases recall@100 by 66% over LightGCN.

WAML outperforms GAT with 2.3x higher recall.

Abstract

We showcase a novel solution to a recommendation system problem where we face a perpetual soft item cold start issue. Our system aims to recommend demanded products to prospective sellers for listing in Amazon stores. These products always have only few interactions thereby giving rise to a perpetual soft item cold start situation. Modern collaborative filtering methods solve cold start using content attributes and exploit the existing implicit signals from warm start items. This approach fails in our use-case since our entire item set faces cold start issue always. Our Product Graph has over 500 Million nodes and over 5 Billion edges which makes training and inference using modern graph algorithms very compute intensive. To overcome these challenges we propose a system which reduces the dataset size and employs an improved modelling technique to reduce storage and compute without loss in performance. Particularly, we reduce our graph size using a filtering technique and then exploit this reduced product graph using Weighted Averaging of Messages over Layers (WAML) algorithm. WAML simplifies training on large graphs and improves over previous methods by reducing compute time to 1/7 of LightGCN and 1/26 of Graph Attention Network (GAT) and increasing recall$@100$ by 66% over LightGCN and 2.3x over GAT.