URecJPQ: Memory-efficient Multimodal Recommendation Models through RecJPQ in Large-Scale Scenarios

2026-06-22

AI Summary

This work addresses the substantial memory overhead in large-scale multimodal recommender systems, which arises from storing full user/item ID embeddings alongside multimodal feature encoders. To mitigate this, the authors propose URecJPQ, the first approach to introduce Joint Product Quantization (JPQ) into multimodal recommendation. URecJPQ replaces fully parameterized embeddings with shared sub-embeddings that are concatenated on-the-fly, achieving significant model compression while preserving or even enhancing recommendation performance. Experimental results on three cross-domain datasets demonstrate that URecJPQ reduces model parameters by 98%–99% and checkpoint size by 86%–98%, with only an average degradation of 8.5% in Recall and 16% in NDCG; notably, it even outperforms baseline models by up to 85% in certain configurations.

Abstract

Training state-of-the-art recommendation models on large-scale industrial datasets can be a challenging task due to the high number of users and items which are typically represented through ID embeddings. Such embeddings typically require a large amount of memory resources, which are not always available. This problem is further exacerbated in multimodal recommendation, in which multimodal item features generally improve recommendation performance, but require more resources to encode. In this paper, we introduce URecJPQ, a Joint Product Quantization method specifically designed for large-scale and multimodal top-k recommendation tasks, in which the vast number of users and items, combined with the available modalities, further increases the memory demands for the computation. The core idea is to represent each user/item not as a fully learned, unique embedding, but rather as a concatenation of shared learned sub-embeddings, thereby significantly reducing the total number of trainable parameters. Our experiments on three widely-used datasets across different domains (movies, baby and sports products) show that URecJPQ can be effectively applied to multimodal recommendation settings. In large scale scenarios, we observe a substantial reduction in checkpoint sizes and the number of trainable parameters (ranging from 86% to 98%, and 98% to 99%, respectively), with only a marginal decrease in accuracy (8.5% on recall and 16% on NDCG, on average), and, in some cases, even performance improvements (up to 85%), as in the baby products domain. Our codebase is available at https://anonymous.4open.science/r/large_mmrecjpq-839B/README.md.

Problem

• multimodal recommendation
• memory efficiency
• large-scale recommendation
• embedding compression
• ID embeddings

Innovation

• Joint Product Quantization
• Memory-efficient Recommendation
• Multimodal Recommendation
• Large-scale Recommendation
• Embedding Compression