TransX improves recommendation speed and accuracy with scalable components
TransX Technical Report: Scaling Sequence Transduction for Recommendation
Information Retrieval
Summary
Recommending things online—like videos or products—needs smart systems that can learn from what you did before. The authors designed TransX, a new approach that breaks this learning into parts that can scale up independently, making it faster and cheaper to run big models. TransX also cleverly reuses work during training and separates tasks during serving to save resources. Their tests showed TransX boosts click and purchase rates and runs much more efficiently than some existing systems. This means better recommendations without extra cost for big online services.
What this means in practice
- •For e-commerce platform engineers: Use TransX to deliver more accurate and cost-efficient personalized product recommendations at scale.
- •For streaming service operators: Integrate TransX to improve content recommendations, increasing user engagement while lowering serving infrastructure costs.
Authors
Da Xu, Liyan Fang, Divya Venugopalan, Sunny Hsu, Xukai Wang, Rishav Roy Chowdhury, Cindy Liang, Nishant Satya Lakshmikanth
Abstract
We present TransX, a sequence-transduction framework and co-designed infrastructure for scaling industrial recommendation with explicit control over model-compute allocation and execution amortization. TransX separates sequence encoding, behavior memory, and candidate-conditioned fusion into independently scalable components. Training amortizes causal trajectory computation across behavioral prefixes and fans out memory states to candidates; serving disaggregates bulk prefill, nearline refresh, and online prediction through compact behavior-memory KV. We observe anisotropic scaling returns: encoder scaling may saturate beyond moderate depth at fixed memory resolution and fusion capacity; increasing supervision shifts the allocation frontier toward candidate-conditioned fusion; and the memory resolution needed to preserve quality remains roughly stable as model capacity grows. We train TransX on click and conversion, post-train its shared behavioral interface for retrieval and preference ranking, and explore generative recommendation offline. In separate production A/B experiments, the core TransX model improves CTR by 6.02% and conversion by 4.45%, while post-trained retrieval and preference-ranking variants improve CTR by 2.96% and conversion by 1.33%, respectively, over corresponding production baselines. Under matched training conditions, the training infrastructure increases supervised-target throughput by 84%. In serving, disaggregation reduces TransX's total serving cost by nearly 80% versus colocated serving at matched traffic and SLO, enabling approximately 5X the per-request model FLOPs of the HSTU-style production baseline at the same total serving cost. At matched per-request model FLOPs, TransX delivers approximately 4X higher request throughput per H200. These results show that online serving cost need not grow proportionally with recommendation model capacity.