Principled Synthetic Data Enables the First Scaling Laws for LLMs in Recommendation

TL;DR

This paper introduces a principled synthetic data generation framework that enables the first reliable scaling laws for large language models in recommendation systems, significantly improving their performance and predictability.

Contribution

It presents a novel layered synthetic data framework that produces high-quality, recommendation-specific data, facilitating robust LLM scaling laws and improved downstream performance.

Findings

Models trained on synthetic data outperform real data-trained models in ranking tasks.

First empirical demonstration of power-law scaling for LLMs in recommendation.

Consistent perplexity reduction across multiple synthetic data modalities.

Abstract

Large Language Models (LLMs) represent a promising frontier for recommender systems, yet their development has been impeded by the absence of predictable scaling laws, which are crucial for guiding research and optimizing resource allocation. We hypothesize that this may be attributed to the inherent noise, bias, and incompleteness of raw user interaction data in prior continual pre-training (CPT) efforts. This paper introduces a novel, layered framework for generating high-quality synthetic data that circumvents such issues by creating a curated, pedagogical curriculum for the LLM. We provide powerful, direct evidence for the utility of our curriculum by showing that standard sequential models trained on our principled synthetic data significantly outperform ($+130\%$ on recall@100 for SasRec) models trained on real data in downstream ranking tasks, demonstrating its superiority for learning generalizable user preference patterns. Building on this, we empirically demonstrate, for the first time, robust power-law scaling for an LLM that is continually pre-trained on our high-quality, recommendation-specific data. Our experiments reveal consistent and predictable perplexity reduction across multiple synthetic data modalities. These findings establish a foundational methodology for reliable scaling LLM capabilities in the recommendation domain, thereby shifting the research focus from mitigating data deficiencies to leveraging high-quality, structured information.