The new era of augmented intelligence includes humans, says IBM engineer | #GuestOfTheWeek


Machine learning first took root in the 1950s and has slowly progressed over six decades through 2010. Since then, IBM and companies such as Google, Microsoft and Facebook have been integrating and improving the technology at a rapid pace. And now many other companies are joining the bandwagon.

On February 23, Apple Inc. announced that it is expanding its artificial intelligence and machine learning team in Seattle. The expanded team will join forces with the company’s headquarters team in Cupertino to develop innovative AI features for future product and services.

“Machine learning is fascinating. It’s over-hyped maybe, but it is also moving very quickly. Every year there is new, cool stuff. Five years ago, nobody spoke about deep learning. Now it is everywhere,” said Jean-Francois Puget, Ph.D. (pictured), distinguished engineer, machine learning and optimization, IBM Analytics, at IBM Corp. Puget has more than 25 years of experience in computer science. He currently works to bring machine learning to the enterprise.

Puget spoke about the IBM Machine Learning platform with Dave Vellante (@dvellante) and Stu Miniman (@stu), co-hosts of theCUBE, SiliconANGLE Media’s mobile live streaming studio, while at the IBM Machine Learning Launch Event in New York, New York. (*Disclosure below.)

This week, theCUBE features Jean-Francois Puget as our guest of the week.

Bringing machine learning to the enterprise

IBM touts that 71 percent of global Fortune 500 companies are on the mainframe and that 80 percent of the world’s corporate data resides or originates there. The announcement of the IBM Machine Learning platform provides an opportunity for the enterprise to tap into the latest technology available on its z Systems mainframe to develop models using historical or recent data to improve business outcomes and detect fraud.

According to IBM, this is “the first cognitive platform for continuously creating, training and deploying a high volume of analytic models in the private cloud at the source of vast corporate data stores.” Puget noted that if everything is on the mainframe, there is a greater advantage to learn from operational data without slowing down your business functions.

Most people equate machine learning with algorithms, but Puget believes it is the steps taken before and after you run algorithms that are the crucial elements for enterprise adoption.

“What do you do with the results of the algorithm is key. … We are focusing on creating value for our customers,” he said.

The IBM platform affords the client ease of use, he added. Although businesses can tap into the open-source community as a resource for algorithms and implement Apache Spark ML, building a platform that delivers valuable insights from data is not an IT function, according to Puget.

Describing the process as a nightmare, Puget illustrated that, first, it is necessary to transform the data to make it usable for machine learning. The next step is to choose the right algorithm; however, he said, moving the models to the right place within the operating system is the major hurdle. “So, our value is to automate what you do before you run the algorithm and then what you do after. That is our differentiator,” Puget stated.

Another added benefit is governance. The IBM Machine Learning platform tracks model creation and gives the customer role-based access control to prevent every employee from deploying models onto the platform.

“So, this solution will come with all the governance and integrity constraints you can expect from us,” Puget said.

Artificial or augmented?

In the battle of human versus machine, the greatest fear is that technology will replace humans. While in some instances that may be true, in other cases IBM thinks of artificial intelligence as a valuable assistant.

“We don’t believe we are here to replace humans; we’re here to assist humans. So, we say augmented intelligence or assistance,” Puget said.

He was adamant that the role of machine learning is to provide additional data to make better decisions and to help businesses become more efficient and productive through machine learning.

IBM Machine Learning is still growing. Puget spoke about upcoming releases for IBM that are also going to support Anaconda on the IBM mainframe so that Python programmers can access big data analytics running on the z/OS platform. Additional future projects include the TensorFlow 0.12 framework, created by Google, Puget said.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE and theCUBE’s coverage of the IBM Machine Learning Launch Event 2017 NYC. (*Disclosure: TheCUBE is a media partner at the conference. Neither IBM nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE