IRLGNov 20, 2024

On-device Content-based Recommendation with Single-shot Embedding Pruning: A Cooperative Game Perspective

arXiv:2411.13052v24 citationsh-index: 13Has CodeWWW
Originality Incremental advance
AI Analysis

This work addresses the storage and computation challenges for deploying recommender systems on resource-constrained devices, such as in federated learning, with an incremental improvement over existing pruning methods.

The paper tackles the storage bottleneck of embedding tables in content-based recommender systems for on-device deployment by proposing Shaver, a method that uses Shapley values to prune embeddings efficiently without retraining, achieving competitive performance across various parameter budgets on three real-world datasets.

Content-based Recommender Systems (CRSs) play a crucial role in shaping user experiences in e-commerce, online advertising, and personalized recommendations. However, due to the vast amount of categorical features, the embedding tables used in CRS models pose a significant storage bottleneck for real-world deployment, especially on resource-constrained devices. To address this problem, various embedding pruning methods have been proposed, but most existing ones require expensive retraining steps for each target parameter budget, leading to enormous computation costs. In reality, this computation cost is a major hurdle in real-world applications with diverse storage requirements, such as federated learning and streaming settings. In this paper, we propose Shapley Value-guided Embedding Reduction (Shaver) as our response. With Shaver, we view the problem from a cooperative game perspective, and quantify each embedding parameter's contribution with Shapley values to facilitate contribution-based parameter pruning. To address the inherently high computation costs of Shapley values, we propose an efficient and unbiased method to estimate Shapley values of a CRS's embedding parameters. Moreover, in the pruning stage, we put forward a field-aware codebook to mitigate the information loss in the traditional zero-out treatment. Through extensive experiments on three real-world datasets, Shaver has demonstrated competitive performance with lightweight recommendation models across various parameter budgets. The source code is available at https://github.com/chenxing1999/shaver

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes