Sep 10

Software AG Launches new Open Source Library for Standards-based AI, Machine Learning

Software AG announced the availability of Nyoka, an open source library that enables data scientists to transform Artificial Intelligence (AI), Predictive Analytics and Machine Learning models into the Predictive Model Markup Language (PMML) industry standard. Nyoka is a Python library that provides comprehensive support for the latest PMML standard as well as extensions for data preprocessing, script execution and deep neural networks. Nyoka underscores Software AG’s commitment to open industry standards and complements its Zementis predictive analytics for enterprise-grade, operational deployment of AI.

Dr. Michael Zeller, Senior Vice President, AI Strategy & Innovation, Software AG, said: “To address the challenges that organizations face with complex AI solutions, frequent model updates, cross-platform execution as well as data integration, Software AG emphasizes a vendor neutral approach that provides users plug-and-play simplicity with a wide range of components. Nyoka streamlines the work of data scientists, reduces the complexities of deploying machine learning models, and gives them more time to focus on creating new models that deliver increased business value.”

PMML is an XML-based predictive model interchange format and the leading standard for statistical and data mining models. The Data Mining Group (DMG), an independent, vendor-led consortium that develops data mining standards, spearheaded the development of PMML for nearly twenty years in support of the Data Science community.

Holger Mueller, VP and Principle Analyst at Constellation Research, said: “The increasingly complex, multi-vendor IT environment, coupled with a rapidly growing collection of open source machine learning packages has multiplied the value of open industry standards. To truly scale smarter, AI-driven applications across the organization, it will be imperative to establish consistent processes that leverage open platforms.”

In its initial release, Nyoka provides a wealth of classes and functions that are designed to make the data scientist’s life easier. For example, Nyoka enables developers to create a PMML file for a Random Forest model from an existing scikit-learn object. Nyoka comes with an extensive HTML documentation and a growing number of Jupyter Notebook tutorials which demonstrate how Nyoka supports the use of PMML as a transport file format for data science models.

With Zementis, Software AG delivers a common predictive analytics strategy across the entire IT ecosystem. In supporting the PMML industry standard, Zementis provides an efficient process for instant operational deployment of models exported through Nyoka as well as many other data science tools, thus addressing the highest execution requirements for batch processing, in-memory computation and streaming data. Zementis is an integral part of the Digital Business Platform and is a strategic part of Software AG’s Cumulocity IoT solution.