Reproduce “Generative Large Recommendation Models: Emerging Trends in LLMs for Recommendation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hao Wang, Wei Guo, Luankang Zhang, Jin Yao Chin, Yufei Ye, Huifeng Guo, Yong Liu, Defu Lian and 2 more
This tutorial outlines the emerging paradigm of Generative Large Recommendation Models (GLRMs), a class of massive-scale recommendation systems designed to obey scaling laws, and details their architectural principles, key challenges, and future research directions. It aims to fill a knowledge gap regarding these large-scale generative models, drawing insights from recent industrial successes with trillion-parameter systems.
Key insights
- Generative Large Recommendation Models (GLRMs) are emerging as a distinct paradigm in recommendation systems, designed at massive scales (e.g., trillions of parameters) and demonstrating powerful scalability.
- The application of scaling laws to recommendation systems is crucial, showing how data and model size influence performance, similar to general-purpose Large Language Models.
- Effective development of GLRMs requires a data-centric approach, emphasizing data quality, sophisticated user behavior modeling (long sequences, multi-behavior, multi-domain), and efficiency optimizations for training and inference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.