IBM Machine Learning

For March 20, 2017 Machine Learning is the cognitive platform for continuous creation, training and deployment of large volume of analytical models in a private cloud.

On March 20, 2017 the IBM company submitted the cognitive IBM Machine Learning platform on the basis of the IBM Watson platform.

According to priorities of the company, a system will become available on z Systems mainframes, operational cores of the global organizations.

Representation of IBM Machine Learning for z/OS, (2017)

IBM Machine Learning allows data processing specialists to automate creation, training and deployment of the operational analytical models supporting:

• any language (for example, Scala, Java, Python);
• any popular framework for machine learning (for example, Apache SparkML, TensorFlow, H2O);
• any data type on transactions;
• data movement in a cloud without additional expenses, delays or risks.

Cognitive Automation for Data Scientists developed by IBM Research helps information processing specialists to select a suitable algorithm for the analysis by comparison of available algorithms with the available data and their ranging. Service also considers different circumstances, for example, necessary functionality of an algorithm and speed of obtaining results.

Machine and deep learning is represented by new spheres of analytics. These technologies will become a basis of automation of process of receiving insights on the scale of the crucial systems and cloud services worldwide. IBM Machine Learning was developed for effective use of key technologies of Watson and acceleration of implementation of machine learning on platforms where an overwhelming part of corporate data is concentrated. As customers notice business return from investments into a private cloud, they will expand application of hybrid and public clouds. Rob Thomas, head of IBM Analytics

The mainframe of IBM z Systems is capable to process to 2.5 billion transactions a day. IBM Machine Learning for z/OS helps to take value from these z Systems, without moving information from a system for the analysis that allows to minimize delays, costs for transactions and risks of security connected with traditional ETL processes. A system constantly analyzes data, models for providing the improved forecasts, instruments of optimization of behavioral models and acceleration of time of receiving insights.

IBM Machine Learning at first will be available on z/OS, then on other platforms, including IBM POWER Systems.

Robotics