NVIDIA at RecSys 2022

Learn about NVIDIA Merlin^™, an open source framework for building recommender systems. Merlin empowers data scientists, machine learning engineers, and researchers to build high-performing recommenders at scale.

In this ACM RecSys 2022 accepted paper, discover how HPS combines a high performance GPU embeddings cache with an hierarchical storage architecture to realize low-latency retrieval of embeddings for online model inference.

Attend this hands-on ACM RecSys 2022 Tutorial and dive into how to make developing and deploying recommender systems easier. Covers methods for evaluating existing approaches, developing new ideas, and deploying them to production.

In this ACM RecSys 2022 accepted submission, learn about NVIDIA Merlin HugeCTR, a framework for click through rate estimation that is optimized for training and inference. It also enables training at scale with model-parallel embeddings and data-parallel neural networks.

With the Merlin Models library, access common architectures that can be coupled with different prediction tasks, loss functions, and negative sampling strategies to ease development and deployment.

During this NVIDIA AI online event, join fellow data scientists and engineers from Netflix, Twitter, NVIDIA, and more to discuss learnings and best practices on building and deploying effective recommender systems.

Introduction to popular Two Tower architecture, why it solves for challenges associated with candidate retrieval, and how using Merlin Models enables training with 4 lines of code.

In this blog, we cover sequential and session-based recommendation tasks, including why it’s important and practical use cases. We also provide a brief overview of our Transformers4Rec solution.

Watch NVIDIA CEO Jensen Huang discuss the importance of AI frameworks, including NVIDIA Merlin™, in the NVIDIA GTC keynote.

Serving user requests with low latency and high accuracy is critical for recommender systems. NVIDIA Merlin introduces HPS, a scalable solution with multilevel adaptive storage to enable deployment of Terabyte-size models under real-time latency constraints!

In this post, learn how NVIDIA Merlin Distributed Embeddings enables developers to rapidly train Terabyte-scale embedding based models in TensorFlow 2 with just a few lines of code! 

Read about the latest trends for recommender system practices within industry. The whitepaper includes insights from industry leads from companies such as Tencent, Meituan, Wayfair, Magazine Luiza, The New York Times, and more.