Last month IBM provided more details about their new Watson Machine Learning Initiatives starting with System z (WMLz). Beginning a new initiative with the mainframe platform is an interesting move for IBM. The new WMLz offerings:
- Introduce an end-to-end enterprise machine-learning platform. Extracting core machine learning technology from IBM Watson, IBM is offering a new platform for creating, training, and deploying analytic models. The new IBM machine-learning platform on the z System mainframe platform supports the operational core for many Fortune 500 international corporations and governments around the world. This allows them to deploy advanced machine learning initiatives using the vast volumes of customer, operational, and other types of data they have collected to achieve a competitive advantage.
- Provide a platform for collaboration with interactive models. Implementing first on System z is optimal since it is where the critical core data resides for today’s machine learning analytics. Machine-learning analytics allow companies to make the best real-time business decisions with the latest data. System z also provides the best platform for collaboration for developing all types of analytical models from all the types of data. Data scientists or analysts working with the System z machine-learning models can collaborate with all of their favorite software programs such as Java, Scala, Python, and many more.
- Provide business processes with cognitive ML workflow capabilities. The new IBM machine-learning capabilities are always gaining knowledge from their ingested data and building their cognitive capabilities. By training and then learning, the machine-learning system continues to build cognitive decision criteria, automating insight at scale for all types of business function and applications as it can take its output as input to continue to learn. During the IBM announcement the USA Cycling Women Team, Wi-Fi provider SolutionInc, and SETI were mentioned as companies already benefiting from machine learning applications.
- Ingest almost any type of data source. IBM machine-learning can ingest any type of data such as transactional, social, Internet of Things, and almost anything else you can think of. From these diverse inputs the machine-learning model can be built and deployed to gain better insights from these and other types of interaction. New and diverse data inputs can be modeled against and with each other and the machine learning model can continuously evaluate the model’s criteria, exposing new opportunities and efficiencies.
- Deploy an automated feedback loop for re-training. Once deployed, IBM’s machine learning models continue to learn as their output is fed back into the model for further learning and model refinements. Constant learning and model refinement help improve the model results through a continuous feedback loop, further refining, retraining, and optimizing the machine learning model.
- RESTful API for imbedding predictive models in any application. IBM machine learning interfaces with a variety of programming languages and APIs. The most common machine learning languages used today by data scientists are Scala, R, Python, and Java. These language APIs are available for incorporation into the machine learning infrastructure now, and more languages are planned to be added in the future.
The new IBM Machine Learning initiatives reinforce IBM commitment to the mainframe as the central core processing server.
Dave Beulke is a system strategist, application architect, and performance expert specializing in Big Data, data warehouses, and high performance internet business solutions. He is an IBM Gold Consultant, Information Champion, President of DAMA-NCR, former President of International DB2 User Group, and frequent speaker at national and international conferences. His architectures, designs, and performance tuning techniques help organization better leverage their information assets, saving millions in processing costs. Follow him on Twitter or connect through LinkedIn.
Leave a Reply