No longer is advanced computing only for the super rich or super intelligent. With thanks to companies like IBM, machine learning is becoming even more accessible. One of their next moves is to bring machine learning to a mainframe near you. Firstly they want to provide a simplified experience for data modeling and model deployment. Secondly, they are looking to support the deployment of machine learning tools on its z series mainframes.
Even though we have fine examples of machine learning networks such as Google’s TensorFlow available, a lot of specialized knowledge is still required in order to run them efficiently. Programs like Apache Spark and Jupyter are the same. While they are very good at what they do, unless you are very computer savvy, you don’t stand a chance at using them to their full potential. IBM’s Machine Learning System, on the other hand, allows data scientists the opportunity to prepare their data for modeling, evaluate various modeling options, choose a modeling type, and deploy their models to IBM’s servers.
Machine Learning was launched by IBM last year on its BlueMix cloud platform and site that it can now support analytic models in any language, in any popular machine learning framework, using any transactional data type. But, now, IBM is looking to deploy these applications in their own private cloud system removing any cost, latency or security worries that come with moving data away from the premises. Argus Health is one company that are benefiting greatly from the likes of IBM. They carry out over 1 billion transactions per year on their IBM z Series mainframes and use the available modeling tools to improve overall standards of care.
There’s no denying that IBM is really pushing forward with these ideas. Currently, the z Series is the only one of IBM’s computers to have Watson’s Machine Learning Tools. But, the company is looking to extend that to some of its other product lines including its POWER platform sometime shortly, so watch this space.
More News To Read