Web recommendations provide personalized items from massive catalogs for users, which rely heavily on retrieval stages to trade off the effectiveness and efficiency of selecting a small relevant set from billion-scale candidates in online digital platforms. As one of the largest Chinese search engine and news feed providers, Baidu resorts to Deep Neural Network (DNN) and graph-based Approximate Nearest Neighbor Search (ANNS) algorithms for accurate relevance estimation and efficient search for relevant items. However, current retrieval at Baidu fails in comprehensive user-item relational understanding due to dissected interaction modeling, and performs inefficiently in large-scale graph-based ANNS because of suboptimal traversal navigation and the GPU computational bottleneck under high concurrency. To this end, we propose a GPU-accelerated Multi-relational Parallel Graph Retrieval (GMP-GR) framework to achieve effective yet efficient retrieval in web-scale recommendations. First, we propose a multi-relational user-item relevance metric learning method that unifies diverse user behaviors through multi-objective optimization and employs a self-covariant loss to enhance pathfinding performance. Second, we develop a hierarchical parallel graph-based ANNS to boost graph retrieval throughput, which conducts breadth-depth-balanced searches on a large-scale item graph and cost-effectively handles irregular neural computation via adaptive aggregation on GPUs. In addition, we integrate system optimization strategies in the deployment of GMP-GR in Baidu. Extensive experiments demonstrate the superiority of GMP-GR in retrieval accuracy and efficiency. Deployed across more than twenty applications at Baidu, GMP-GR serves hundreds of millions of users with a throughput exceeding one hundred million requests per second.

Deep Neural Network (DNN) based recommendation systems are widely used in the modern internet industry for a variety of services. However, the rapid expansion of application scenarios and the explosive global internet traffic growth have caused the industry to face increasing challenges to serve the complicated recommendation workflow regarding online recommendation efficiency and compute resource overhead. In this paper, we present a GPU-accelerated online serving system, namely Lion, which consists of the staged event-driven heterogeneous pipeline, unified memory manager, and automatic execution optimizer to handle web-scale traffic in a real-time and cost-effective way. Moreover, Lion provides a heterogeneous template library to enable fast development and migration for diverse in-house web-scale recommendation systems without requiring knowledge of heterogeneous programming. The system is currently deployed at Baidu, supporting over twenty recommendation services, including news feed, short video clips, and the search engine. Extensive experimental studies on five real-world deployed online recommendation services demonstrate the superiority of the proposed GPU-accelerated online serving system. Since launched in early 2020, Lion has answered billions of recommendation requests per day, and has helped Baidu successfully save millions of U.S. dollars in hardware and utility costs per year.

In modern internet industries, deep learning based recommender systems have became an indispensable building block for a wide spectrum of applications, such as search engine, news feed, and short video clips. However, it remains challenging to carry the well-trained deep models for online real-time inference serving, with respect to the time-varying web-scale traffics from billions of users, in a cost-effective manner. In this work, we present JIZHI - a Model-as-a-Service system - that per second handles hundreds of millions of online inference requests to huge deep models with more than trillions of sparse parameters, for over twenty real-time recommendation services at Baidu, Inc. In JIZHI, the inference workflow of every recommendation request is transformed to a Staged Event-Driven Pipeline (SEDP), where each node in the pipeline refers to a staged computation or I/O intensive task processor. With traffics of real-time inference requests arrived, each modularized processor can be run in a fully asynchronized way and managed separately. Besides, JIZHI introduces the heterogeneous and hierarchical storage to further accelerate the online inference process by reducing unnecessary computations and potential data access latency induced by ultra-sparse model parameters. Moreover, an intelligent resource manager has been deployed to maximize the throughput of JIZHI over the shared infrastructure by searching the optimal resource allocation plan from historical logs and fine-tuning the load shedding policies over intermediate system feedback. Extensive experiments have been done to demonstrate the advantages of JIZHI from the perspectives of end-to-end service latency, system-wide throughput, and resource consumption. Since launched in July 2019, JIZHI has helped Baidu saved more than ten million US dollars in hardware and utility costs per year while handling 200% more traffics without sacrificing the inference efficiency.