Covering Disruptive Technology Powering Business in The Digital Age

SAS is embedding more artificial intelligence (AI) and automation in the SAS Platform, making it easier for customers to build AI solutions based on proven and trusted analytics. The latest SAS^® Viya^® release also includes significant enhancements designed to maximize collaboration among data scientists and business users, and to improve the creation and deployment of powerful analytics models.

“In today’s analytics economy, enterprises require a powerful analytics platform with end-to-end capabilities, from data preparation and data management to creation and deployment of models,” said Randy Guard, Chief Marketing Officer at SAS. “The SAS Platform and its SAS Viya capabilities provide an open foundation from which data scientists, business users and executives can transform raw data into intelligence.”

What’s New?
Among the enhancements available this June:

Embedded AI and automation. SAS continues to embed AI capabilities and enable automation across the SAS Platform, from data preparation and model building to model deployment. Good data is the bedrock for good analysis, and the new SAS Viya capabilities help data scientists and business users clean, prepare and transform data for analysis.

By automating data prep, users can identify the most significant variables and fields for more focused and effective analysis. SAS Viya also automates model building and deployment, so various models can be tested and the best ones quickly chosen. Since the key to good models is feature engineering, SAS will add new feature engineering capabilities to improve model accuracy.

For natural language processing, SAS is automating sentiment analysis and document classification using recurrent neural networks (RNN).

More transparency. With more advanced machine learning and deep learning methods, it is imperative to support the principles of fairness, accountability and transparency in AI and machine learning. The latest SAS Viya release does just that, offering improved interpretability and transparency through use of industry frameworks such as Local Interpretable Model-Agnostic Explanations (LIME) and impact, confidence and ease (ICE).

“Users should not be forced to rely on a ‘black box’ when using AI and machine learning techniques,” said Saurabh Gupta, Director of Analytics Products at SAS. “By combining SAS AI and advanced analytics with industry frameworks like LIME and ICE, data scientists and business users can more easily interpret and explain the logic behind a model’s predictions.”

Better data governance. One of the hallmarks of this latest SAS Platform release is improved data lineage across the entire analytics life cycle. Visualizing the relationships between data – both structured and unstructured – builds an intuitive understanding of how data depends on other data. With this knowledge, it becomes easier to find errors and identify redundant copies of data, ensuring you are using the right one.

In addition to improved data governance, SAS Viya offers model performance dashboards, helping users to easily see with a click how each of their models is performing. Dashboards also give the option of scheduling or automating model updates when new data has been added or when it is underperforming against expectations.

Improved user experience. To make analytics pervasive in an organization, SAS continues to focus on delivering a consistent user experience for data scientists, analysts and developers. Designed to improve productivity and enable collaboration, the interface in the new enhancements helps users seamlessly transition through all stages of the analytics life cycle – from data to discovery to deployment. To further foster collaboration, SAS Drive serves as a central repository for projects across SAS products.

Additional enhancements to the user experience include improved GPU support and operating system support for SUSE Linux and Windows symmetric multiprocessing (SMP).

More open. The enhanced SAS Platform adds capabilities that improve its openness and interoperability with other systems. The release enables users to embed open source code within their SAS machine learning pipelines. The Platform also includes open access to Salesforce and Java Database Connectivity (JDBC), and the ability to include open source JavaScript visualizations in explorations and reports. App developers are also equipped with mobile software development kits (SDKs) to easily build a variety of customized and personalized mobile apps for smartphones and tablets. The release also provides Python model support in decision flows.

(0)(0)