A Decoding Acceleration Framework for Industrial Deployable LLM-based Recommender Systems

Recently, increasing attention has been paid to LLM-based recommender systems, but their deployment is still under exploration in the industry. Most deployments utilize LLMs as feature enhancers, generating augmentation knowledge in the offline stage. However, in recommendation scenarios, involving...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-08
Main Authors: Yunjia Xi, Wang, Hangyu, Chen, Bo, Lin, Jianghao, Zhu, Menghui, Liu, Weiwen, Tang, Ruiming, Zhang, Weinan, Yu, Yong
Format: Article
Language:English
Subjects:
Acceleration
Recommender systems
Retrieval
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Recently, increasing attention has been paid to LLM-based recommender systems, but their deployment is still under exploration in the industry. Most deployments utilize LLMs as feature enhancers, generating augmentation knowledge in the offline stage. However, in recommendation scenarios, involving numerous users and items, even offline generation with LLMs consumes considerable time and resources. This generation inefficiency stems from the autoregressive nature of LLMs, and a promising direction for acceleration is speculative decoding, a Draft-then-Verify paradigm that increases the number of generated tokens per decoding step. In this paper, we first identify that recommendation knowledge generation is suitable for retrieval-based speculative decoding. Then, we discern two characteristics: (1) extensive items and users in RSs bring retrieval inefficiency, and (2) RSs exhibit high diversity tolerance for text generated by LLMs. Based on the above insights, we propose a Decoding Acceleration Framework for LLM-based Recommendation (dubbed DARE), with Customized Retrieval Pool to improve retrieval efficiency and Relaxed Verification to increase the acceptance rate of draft tokens, respectively. Extensive experiments demonstrate that DARE achieves a 3-5x speedup and is compatible with various frameworks and backbone LLMs. DARE has also been deployed to online advertising scenarios within a large-scale commercial environment, achieving a 3.45x speedup while maintaining the downstream performance.
ISSN:2331-8422