Balanced Co-Clustering of Users and Items for Embedding Table Compression in Recommender Systems

2026-04-20

AI Summary

This work addresses the challenge of excessive parameter count in embedding tables of industrial-scale recommender systems, which incurs substantial computational and memory costs. Existing compression methods often compromise recommendation accuracy or introduce additional computational overhead. To overcome these limitations, the authors propose BACO, a novel framework that integrates graph clustering theory into embedding compression. BACO formulates a balanced co-clustering objective leveraging collaborative signals from user–item interactions and employs a weighted label propagation solver coupled with a two-stage user clustering mechanism to effectively prevent codebook collapse. Evaluated on multiple benchmark datasets, BACO achieves embedding compression rates exceeding 75% with a maximum recall degradation of only 1.85%, while accelerating training by up to 346× compared to the strongest baseline.

Technology Category

• Data Mining & Knowledge Management: Data Compression
• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Combinatorial Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• User Modeling, Personalization and Recommendation: Social recommender systems and personalization for the social Web
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text

Abstract

Recommender systems have advanced markedly over the past decade by transforming each user/item into a dense embedding vector with deep learning models. At industrial scale, embedding tables constituted by such vectors of all users/items demand a vast amount of parameters and impose heavy compute and memory overhead during training and inference, hindering model deployment under resource constraints. Existing solutions towards embedding compression either suffer from severely compromised recommendation accuracy or incur considerable computational costs. To mitigate these issues, this paper presents BACO, a fast and effective framework for compressing embedding tables. Unlike traditional ID hashing, BACO is built on the idea of exploiting collaborative signals in user-item interactions for user and item groupings, such that similar users/items share the same embeddings in the codebook. Specifically, we formulate a balanced co-clustering objective that maximizes intra-cluster connectivity while enforcing cluster-volume balance, and unify canonical graph clustering techniques into the framework through rigorous theoretical analyses. To produce effective groupings while averting codebook collapse, BACO instantiates this framework with a principled weighting scheme for users and items, an efficient label propagation solver, as well as secondary user clusters. Our extensive experiments comparing BACO against full models and 18 baselines over benchmark datasets demonstrate that BACO cuts embedding parameters by over 75% with a drop of at most 1.85% in recall, while surpassing the strongest baselines by being up to 346X faster.

Problem

• embedding compression
• recommender systems
• embedding tables
• parameter efficiency
• resource constraints

Innovation

• balanced co-clustering
• embedding compression
• collaborative signals
• codebook sharing
• label propagation