5 papers
SilverTorch: A Unified Model-based System to Democratize Large-Scale Recommendation on GPUs
Bi Xue, Hong Wu, Lei Chen +29
Serving deep learning based recommendation models (DLRM) at scale is challenging. Existing approaches rely on dedicated ANN indexing and filtering services on CPUs, suffering from…
FreeScale: Distributed Training for Sequence Recommendation Models with Minimal Scaling Cost
Chenhao Feng, Haoli Zhang, Shakhzod Ali-Zade +17
Modern industrial Deep Learning Recommendation Models typically extract user preferences through the analysis of sequential interaction histories, subsequently generating predictio…
LIME: Link-based user-item Interaction Modeling with decoupled xor attention for Efficient test time scaling
Yunjiang Jiang, Ayush Agarwal, Yang Liu +1
Scaling large recommendation systems requires advancing three major frontiers: processing longer user histories, expanding candidate sets, and increasing model capacity. While prom…
MoEBlaze: Breaking the Memory Wall for Efficient MoE Training on Modern GPUs
Jiyuan Zhang, Yining Liu, Siqi Yan +6
The pervasive "memory wall" bottleneck is significantly amplified in modern large-scale Mixture-of-Experts (MoE) architectures. MoE's inherent architectural sparsity leads to spars…
Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment
Dai Li, Kevin Course, Wei Li +13
Scaling laws have been established for recommender systems, yet efficiently deploying foundation model (FM) across multiple recommendation surfaces remains a major unsolved challen…