O’REILLY、INTEL AI 主办
Put AI to work
2018年4月10-11日:培训
2018年4月11-13日:辅导课 & 会议
北京,中国

Developing artificial intelligence applications in the cloud for time to deployment (TTD)

This will be presented in English.

Le Zhang (Microsoft), Graham Williams (Microsoft)
14:5015:30 Friday, April 13, 2018

必要预备知识 (Prerequisite Knowledge)

云计算以及R和Python等编程语言中级知识。 机器学习和分布式大数据平台(如Spark)高级知识。

Intermediate knowledge of cloud computing, R, and Python, and advanced knowledge of machine learning and distributed big data platform (e.g., Spark).

您将学到什么 (What you'll learn)

如何利用微软云平台组件和服务(如数据科学/深度学习虚拟机、机器学习服务等)开发和部署一个企业级端到端人工智能应用(基于Spark的可扩展推荐系统)。

Learn how to develop and operationalize end-to-end enterprise-grade artificial intelligence applications for real-world scalable Spark-based recommender system on Azure cloud.

描述 (Description)

本讲话将用英语授课,同时会提供中文同声传译。中文版本摘要会在英文摘要下面给出。

Contemporary AI and advanced analytical solution development requires multiple iterations, and complexity of the iterative process may grow exponentially if there is lack of elastic computing environment and collaborative working platform. Considering timely productization of AI models, time-to-deployment is pivotal to enterprise. To this end, a set of tools and methodology for data scientists/engineers and AI developers to easily develop AI applications is extremely important. The talk introduces how data scientists and AI developers can be empowered with Azure cloud services (e.g., Azure Machine Learning Services, Team Data Science Process, Data Science Virtual Machine, etc.) and open source toolkits (e.g., Spark) for convenient collaborative experimentation and production for building their AI applications. A walk-through for building a Spark-based movie recommender system will be demonstrated, to illustrate how the introduced Azure products and services are used for achieving such solution deployment.

人工智能和高级数据科学分析应用的开发往往需要多次迭代,同时其开发复杂度会随数据量增大、模型增多以及参与人员人数增多而指数级增加。基于成本和上线速率的考虑,”部署时效性”(Time-To-Deployment)对企业级人工智能应用的开发至关重要,因而一套可供数据科学家、人工智能应用及机器学习工程师,以及商业决策人协同合作的平台及方法将为人工智能开发快速迭代提供极大便利。本讲将从实例应用出发,介绍如何基于微软Azure云平台,利用Azure机器学习服务、团队数据科学进程(Team Data Science Process)框架,以及常用开源工具(如Spark等)开展高效的企业级人工智能应用开发,以应对”部署时效性”的要求。本讲将通过一个基于Spark的电影推荐系统实例来展示如何使用这些工具来搭建一个完整的解决方案。

Photo of Le Zhang

Le Zhang

Microsoft

Le Zhang is a data scientist with Microsoft Cloud and Artificial Intelligence, where he applies cutting-edge machine learning and artificial intelligence technology to accelerate digital transformation for enterprises and startups on cloud. He’s helped numerous corporations develop and build enterprise-grade scalable advanced data analytical systems with a broad spectrum of application scenarios like manufacturing, predictive maintenance, financial services, ecommerce, and human resource analytics. Le specializes in cloud computing, big data technologies, and artificial intelligence. He enjoys sharing knowledge and learning from people and is a frequent speaker at industrial and academic conferences and community meetups. He holds a PhD in computer engineering.

Photo of Graham Williams

Graham Williams

Microsoft

Graham Williams is director of data science at Microsoft, where he is responsible for the Asia-Pacific region, an adjunct professor with the University of Canberra and the Australian National University, and an international visiting professor with the Chinese Academy of Sciences. Graham has 30 years’ experience as a data scientist leading research and deployments in artificial intelligence, machine learning, data mining, and analytics. Previously, he was principal data scientist with the Australian Taxation Office and lead data scientist with the Australian Government’s Centre of Excellence in Analytics, where he assisted numerous government departments and Australian industry in creating and building data science capabilities. He has also worked on many projects focused on delivering solutions and applications driven by data using machine learning and artificial intelligence technologies. Graham has authored a number of books introducing data mining and machine learning using the R statistical software.