Sign up now: July 29 recommendation system summit 2022

Welcome to come from NVIDIA、 Alibaba 、 tencent 、 Meituan and other experts , Join by NVIDIA Sponsored recommendation system summit .

7 month 29 Day by NVIDIA Sponsored 【 Recommendation system summit 2022】 Coming next week ！ Welcome and NVIDIA、 Alibaba 、 tencent 、 Experts from meituan and other companies communicate online , Understand the relevant knowledge and best practices of building and deploying an effective modern recommendation system .

Whether you are a relevant practitioner 、 Data scientist 、 Researchers or engineers , All need to deal with and build 、 Training 、 Challenges associated with optimizing and deploying production ready recommender systems . traditional method （ For example, matrix decomposition ） Can alleviate to some extent Providing relevant suggestions and personalized interaction are deeply challenging . However , As more subtle contextual data increases , Modern recommendation systems can take advantage of technologies other than building user item matrices 、 Methods and algorithms . Delve into the nuances of recommendation systems , And use a series of tools 、 Method 、 Software packages and libraries , It can help us fine tune and expand our work .

Four summit agendas Six specially invited experts

• NVIDIA Merlin GPU Introduction to the recommendation system

Gao Wenwen NVIDIA Senior Product Manager

Wang zehuan NVIDIA Senior Development Manager

Merlin yes NVIDIA stay GPU Top to end recommendation system framework , Deal with before , Training and reasoning provide optimization and acceleration solutions . This presentation will focus on the following issues ：

1. High performance ： How to cross platform to achieve up to 12 Training times faster

2. Economy is applicable ： How to make the best use of everything , Combined with the characteristics of business model and hardware cost , Achieve efficient training and deployment TB Large scale model

3. Flexible and easy to use ： How to experiment with various models in just a few lines of code

• Search for GPU Training and Forecasting

Liu tongxuan Alibaba cloud intelligent computing platform division PAI Senior technical expert

DeepRec (PAI-TF) It is the unified large-scale sparse model training of Alibaba group / Prediction engine , The lecturer will introduce you DeepRec in GPU Optimize related work , Include ：

1. GPU Runtime Optimize

2. EmbeddingVariable Multilevel hybrid storage

3. Distributed optimization - HB & SOK with EmbeddingVariable

• Meituan local life scene large-scale recommendation system GPU practice

Let Jiaheng Meituan machine learning engine technology expert
Chen Zhuo Meituan machine learning engine technology expert

Recommend system training scenarios in meituan , As the model becomes more and more complex ,CPU The marginal effect of optimization is getting lower and lower . Meituan is based on in-depth customization NVIDIA TensorFlow、HugeCTR, Research and development Booster GPU Training structure , It only takes one line of code to complete GPU Schema migration .

• tencent PCG Large scale machine learning framework - Immeasurable GPU Training accelerates

Luo Zhaokai tencent PCG Machine learning engine technologist

GPU Accelerated training and reasoning prediction have become the current “ Search and push ” Direction is a hot research direction of machine learning framework . This sharing will focus on Tencent PCG Large scale machine learning framework - Immeasurable is GPU Landing and practical experience in training acceleration , discuss GPU High cost performance solution for the implementation of heterogeneous Acceleration Technology .

Summit time and guests

date ：2022 year 7 month 29 Japan Friday

Time : Afternoon 2: 00 - 4: 30｜ Beijing time.

Guest speaker

• NVIDIA、 Alibaba 、 tencent 、 Meituan

Registration method

• Click to register now ：https://marketing.csdn.net/p/b271c49beaf3e15016d5aa31df94ffcd

Or scan the QR code of the poster below to register ：