A powerful cognitive and deep learning tool

Share this post:

From time to time, we invite industry thought leaders to share their opinions and insights on current technology trends to the In The Making blog. The opinions in these blogs are their own, and do not necessarily reflect the views of IBM.

Getting the right tool for the job is essential for anything from home improvement projects to launching satellites. I view the new trend of applying AI, deep learning and cognitive techniques to enterprise IT solutions as following that basic principle. Some tools are more complex and difficult to create than others, but they should all be viewed as a means to an end, not the end in itself.

A recent, stunning announcement by IBM is a great example of what I mean. Calling its new invention “the jet engine of deep learning,” the company recently announced its Distributed Deep Learning (DDL) library for PowerAI, which hooks into TensorFlow (an open-source environment originally developed at Google), and other deep-learning frameworks such as Caffe, Torch, and Chainer.

I previously wrote about a project involving some of my colleagues using deep learning with IBM Power Systems to examine thousands of radiological images in 0.03 percent of the time it takes radiologists to perform the task. The platform they’re using is TensorFlow. It is great to see IBM take this next step and formally integrate it and other world-changing technologies into the DDL library.

Leveraging powerful systems

Archimedes wrote that he could move the entire earth with the proper lever. He would no doubt like the deep-learning leverage contained within the DDL. How powerful is it? According to IBM, it can scale tens of servers and hundreds of GPUs.

The company cites a program in which it incorporated 64 IBM Power Systems servers and 256 NVIDIA Tesla P100 with NVLink GPU accelerators to train the ResNet-101 visual recognition system on the ImageNet-22K data set, shrinking execution time from 16 days to seven hours. That’s a time reduction of slightly more than 98 percent for those of you keeping score.

Technical details about ResNet-101 can be found within its github community. The ImageNet project is a massive image collection designed to aid researchers doing visual recognition projects.

The practical effects of this sort of research in general, and IBM’s performance breakthrough in particular, should extend through several economic sectors. Although I enjoy the demonstrations of an AI-driven system recognizing a cat as much as anyone, it’s nice to see learning and cognitive research developing next steps.

The DDL project has the promise to be applied to use cases in property security, connected cars and traffic grids, robotics (particularly the use of robots in dangerous environments), and other instances where a strong set of eyes coupled with cognition can come in handy. The implication is that such systems will also want to act upon what they perceive. It appears IBM is releasing more toolchain for these inference workloads–with a new PowerAI Vision Technology Preview.

Be careful what you wish for

This should lead to discussions about the good uses to which these tools can be put, as well as the potential downsides. The Industrial Age and the more recent Nuclear Age have shown humankind’s gift for creating powerful tools, coupled with a difficulty — even inability — to control them.

The emerging Machine Learning Age will be no different. As researchers create ever-more-powerful cognitive tools, we must realize we do have the means to control our systems and put them to use for the good of us all.

More stories

IBM Z defines the future of hybrid cloud

Cloud computing, Ecosystem & partners, Mainframes

 The stakes remain high for IT leaders. Not only do they have to respond to dynamic business demands in real time, they must optimize operations on their increasingly complicated multi-platform, multicloud environments– all while simultaneously managing resources and budgets. In short, the mandate for IT leaders remains: “Do more with less.” Today, IBM is announcing more

Open innovation accelerated with Red Hat Enterprise Linux 8

Linux on Power Systems, Power servers, Power Systems

On IBM Power Systems, IBM Z and LinuxONE For more than 18 years, IBM and Red Hat have had a strong history of collaboration, including delivering the world’s two most powerful supercomputers, the US Dept of Energy’s Summit and Sierra supercomputers. This partnership has been a testament to our shared values around open source development more

Application modernization with IBM Cloud Private on IBM Power Systems

Cloud computing, IBM Systems Lab Services, Power Systems

Leading companies are realizing that cloud is how to deploy IT, not where. Cloud capabilities have become a prerequisite for consuming IT resources, and organizations are now focusing on modernization of applications with a cloud native approach. You have many approaches for moving into the cloud, including: Modernizing and optimizing your existing on-premises applications Creating more