Token Factory: Efficiently Integrating Diverse Signals into Large Recommendation Models
For practitioners deploying large recommendation models, Token Factory offers an efficient way to incorporate heterogeneous signals without excessive computational cost.
Token Factory integrates diverse traditional signals into large recommendation models as soft tokens, preventing prompt length explosion and reducing memory/computation overhead while improving performance in a production-scale environment.
Large Recommendation Models (LRMs) have demonstrated promising capabilities in industry-scale recommendation tasks. However, holistically integrating traditional signals into these transformer-based architectures effectively and efficiently remains a major challenge. Conventional approaches that "textualize" these signals directly or create discrete item representations often lead to excessively long prompts, substantial memory footprints, and high computational overhead. To overcome these limitations, we propose "Token Factory", a framework designed to transform traditional signals into "soft tokens" that can be directly processed by LRMs. This approach enables efficient integration and compression of heterogeneous input features, preventing prompt length explosion while enhancing model performance. We detail the architecture of Token Factory and present experimental results validating its effectiveness in a production-scale recommendation environment.