Since the launch of Amazon sagemaker in 2017, we have made great progress, as evidenced by the increasing number of customers using the service. However, due to the relative immaturity of ML tools, ML development workflow is still very repetitive, and management is challenging for developers. Many tools (debugger, project management, collaboration, monitoring, etc.) that developers take for granted when building traditional software have been invented for ml.
For example, when trying a new algorithm or adjusting super parameters, developers and data scientists usually conduct thousands of experiments on Amazon sagemaker, and they need to manually manage all experiments. With the passage of time, it becomes more and more difficult to track the best performance model and use the lessons learned in the experimental process.
Amazon sagemaker studio finally unified all the tools needed for ML development. Developers can write code, track experiments, visualize data, debug and monitor in an integrated visual interface, which greatly improves the work efficiency of developers.
In addition, because all these steps of ML workflow are tracked in the environment, developers can quickly move back and forth between steps, and clone, adjust and replay them. This enables developers to make changes quickly, observe results, and iterate faster, reducing time to market for high-quality ml solutions.
Amazon sagemaker studio enables you to manage the entire ml workflow through a single pane. Let me bring you a whirlwind journey!
With Amazon sagemaker notebooks (currently in preview), you can enjoy an enhanced notebook experience, making it easy to create and share jupyter notebooks. You can quickly switch from one hardware configuration to another without managing any infrastructure.
With Amazon sagemaker experiments, you can organize, track and compare thousands of ML jobs: training jobs or data processing and model evaluation jobs run using Amazon sagemaker processing.
With Amazon sagemaker debugger, you can debug and analyze complex training problems and receive alerts. It will automatically check your model, collect debugging data and analyze it to provide real-time alerts and suggestions, so as to optimize training time and improve model quality. When training the model, all information is visible.
With Amazon sagemaker model monitor, you can detect quality deviations of deployed models and receive alerts. You can easily visualize problems that may affect the model, such as data drift. No code required: just a few clicks.
With Amazon sagemaker autopilot, you can automatically build models under complete control and visibility. Algorithm selection, data preprocessing and model tuning, as well as all infrastructure, will be performed automatically.
Thanks to these new capabilities, Amazon sagemaker now covers a complete ml workflow to build, train, and deploy machine learning models quickly and on any scale.
On April 30, Amazon sagemaker was launched in AWS China (Beijing) operated by halo new network and AWS China (Ningxia) operated by Xiyun data.
How can it help Chinese enterprises drive their business development through machine learning? Let’s see how TA does it~
Jiayi mutual Entertainment
By using the server free architecture provided by AWS to build game platform, big data and machine learning solutions to create user services for thousands of people, Jiayi mutual entertainment has achieved a series of significant benefits.
Chengdu Jiayi mutual Entertainment Technology Co., Ltd. (hereinafter referred to as “Jiayi mutual entertainment”) is a game enterprise focusing on the overseas leisure game market and integrating product R & D, operation and distribution. The company is headquartered in Chengdu, China, and has branches in the United States, Hong Kong and other places. Since its establishment three years ago, the company has been focusing on creating high-quality games and continuously exploring the market potential. So far, the cumulative overseas downloads of the company’s self-developed products have exceeded 110 million person times, of which not only a number of games have been recommended by apple and Google stores worldwide, but also a variety of different types of products, ranking at the top of the download list and classified best seller list in key countries around the world.
In the view of Jiayi mutual entertainment, in the process of transformation from traditional stand-alone games to online games, it is a wise choice to rely on the cloud platform to realize the deployment of services and rapid overseas expansion. A reliable cloud service provider can provide a continuous, stable and non perceived service experience. After the trial evaluation of the mainstream cloud service providers in the market, after a comprehensive investigation, Jiayi mutual entertainment finally chose Amazon Web services.
In order to realize the classification and personalized service for thousands of users, Jiayi mutual entertainment adopted the traditional manual statistical analysis combined with ab test to carry out extensive intelligent marketing. Zhang Yunong said frankly: “machine learning is something we always want to do but can’t do. In order to speed up the launch of new games and seize market opportunities, the team must focus on game research and development, and machine learning technology also needs professionals and strong technical investment. These are the obstacles we face.” AWS has an experienced professional service team to help Jiayi mutual entertainment overcome this difficulty. It not only provides strong technical support and guidance, so that Jiayi mutual entertainment can use Amazon sagemaker to build a set of machine learning prediction platform, make full use of various data in the game, provide a scientific model for payment and retention prediction, and improve the overall business income. At the same time, with the help of AWS professional service team, Jiayi mutual Entertainment Group has established a professional team to be responsible for the research of machine learning / artificial intelligence. In the process of learning and using AWS, the technical vision of team members has been continuously broadened, the development thinking has been gradually upgraded, and the technical experience and ability of AWS have been empowered to team members.
Amazon sagemaker now covers a complete ml workflow to build, train, and deploy machine learning models quickly and on any scale.
Want to get started with Amazon sagemaker’s children’s boots, look! Service deduction voucher direct charging welfare, waiting for you! Value added benefits
From April 30, 2020 to September 30, 2020, AWS provides exclusive benefits for machine learning applications, and 1000 yuan / US $200 service deduction voucher is directly charged to AWS account. I wish you an easy experience of Amazon sagemaker service
You can't miss the time and don't come again. Click the link to get it immediately~