Google launches AI Platform Prediction as generally availability, AI Platform comes with perimeter around machine learning models feature.
USA: Mon, 28 Sep 2020 17:31, by: Deepak Kumar

Search Engine and Technology giant, Google LLC today announced the launch of its AI Platform Prediction as GA (General Available). This platform will enable the companies to host their machine learning model on Google’s public cloud, where infrastructure management will be handled automatically by Google. The AI Platform Prediction comes with the machine learning infrastructure and tools for fast deployment of AI models without hassle of setting up of the complete machine learning infrastructure.

The AI Platform Prediction lets developers in preparing, building, running and sharing machine learning models in the highly scalable cloud environment. This environment is highly secured and can be used for running the production applications. The AI Platform Prediction of Google is based on the Google Kubernetes Engine and the system is designed well for high reliability, flexibility, low overhead latency and high performance of any type of AI workload.

Companies and research institutes looking for highly scalable AI platform can use the Google’s AI Platform Prediction for running their AI workloads. This platform can be configured to scale based on the workload requirements. The newly launched platform is very flexible and can be set-up with less effort. Usually setting-up of machine learning infrastructure takes a lot of effort, while this platform is fully automated and can be configured fast to meet the business needs.

Companies can setup and use this platform without having to worry about infrastructure management and maintenance. The process of setting up a production-grade machine learning and AI environment is highly technical work. This requires lot of experience and time to set-up fully functional machine learning environment. Even large enterprise finds many difficulties in set-up of machine learning and AI platform. So, the Google’s AI Platform Prediction will help companies and research institutes to use the service to run their applications fast.

Google’s AI Platform Prediction is based on the Google’s GKE managed Kubernetes service, which allows the companies to use the platform without having set up and maintenance of the environment by own. This platform will receive several new features and improvements as part of future releases. This makes the system very helpful for the business owners as they will be able to use the latest in Data Science and Artificial Intelligence in their applications.

A new feature so-called perimeter around machine learning models is added to the current release of AI Platform Prediction, which enables the companies to isolate their models from the rest of the company’s cloud environment. This is one of the most advanced feature for protecting the model and data in the cloud hosting environment where large number of companies are running their applications. This further improves the security of the Google’s AI Platform Prediction. Such type of isolation is very helpful in the event of breach of security when hackers move deeper into the system by hoping from one application to another.

The latest version of AI Platform Prediction brings another new feature called Resource Metrics, which is a monitoring tool and administrators can view the utilization of the cloud infrastructure. This also helps in identifying the optimization opportunities to lower the cost of cloud usages. This way administrator can use the monitoring console for lowering the hardware costs and optimize the overall performance of their cloud system.

Data Scientists around the world are using different artificial intelligence techniques to solve unique business problems and most of the time many different development frameworks are used. To extend the wide technologies support AI Platform Prediction comes with the many AI libraries support. Google added the support for the XGBoost and scikit frameworks on its AI Platform Prediction. So, applications developed in these technologies can run on the Google’s AI Platform Prediction.

Google engineers Bhupesh Chandra and Robbie Haertel said in this blogpost “AI Platform makes it simple to deploy models trained using these frameworks with just a few clicks — we’ll handle the complexity of the serving infrastructure on the hardware of your choice.”

Google Cloud product manager Morgan McLean further said “With as little as one command, you can create a policy that governs existing and new VMs, ensuring proper installation and optional auto-upgrade of both agents.”

Google’s of AI Platform Prediction also comes with the general-purpose monitoring features, which helps the administrators in tracking the health of cloud infrastructure. Google also offers agent application which can be installed to track the metrics such as CPU, disk and memory usages.

The Google’s AI Platform Prediction can be used to host models and run the predictions using the model in the cloud environment. In machine learning the process of hosting a model for prediction in production environment is called deployment.

The Google’s AI Platform Prediction is designed to manage the infrastructure needed to run pre-trained model at scale in the production environment. The deployed model can be used both for online and batch prediction requests.

Ads: Do you want to learn Data Science in detail then check Data Science - Guide to Data science, machine learning, deep learning and artificial intelligence. 


Deepak Kumar - Editor

Deepak Kumar is Science Graduate from Delhi University with more than 17 years of experience in Media and Technical field. With a background in Science and Media field, Deepak has been offering services in the media houses. He has worked as director and chief in many companies. During his journalistic career, he has achieved many milestones working in various organisations. As a technical writer, he is of scientific bend of mind and has done many glorious stories in the field during the career.

Address: D-16/116, Sector-3, Rohini Delhi - 110085 India


Phone: +91 9313063554