Skip to main content

IBM is cutting deep-learning processing times from days down to hours

deep learning
Daniel Kaesler/123RF
Deep learning uses algorithms inspired by the way human brains operate to put computers to work on tasks too big for organic gray matter. On Monday, IBM announced that a new record for the performance of a large neural network working with a large data set.

The company’s new deep-learning software brings together more than 256 graphics processing units across 64 IBM Power systems. The speed improvements brought about by the research come as a result of better communication between the array of GPUs.

Recommended Videos

Faster GPUs provide the necessary muscle to take on the kind of large scale problems today’s deep-learning systems are capable of tackling. However, the faster the components are, the more difficult it is to ensure that they are all working together as one cohesive unit.

As individual GPUs work on a particular problem, they share their learning with the other processors that make up the system. Conventional software is not capable of keeping up with the speed of current GPU technology, which means that time is wasted as they wait around for one another’s results.

Hillery Hunter, IBM’s director of systems acceleration and memory, compared the situation to the well-known parable of the blind men and the elephant. The company’s distributed deep-learning project has resulted in an API that developers can be used in conjunction with deep-learning frameworks to scale to multiple servers, making sure that their GPUs remain synchronized.

IBM recorded image recognition accuracy of 33.8 percent on a test run using 7.5 million images from the ImageNet-22K database. The previous best-published result was 29.8 percent, which was posted by Microsoft in October 2014 — in the past, accuracy has typically edged forward at a rate of about one percent in new implementations, so an improvement of four percent is considered to be a very good result.

Crucially, IBM’s system managed to achieve this in seven hours; the process that allowed Microsoft to set the previous record took 10 days to complete.

“Speed and scalability, which means higher accuracy, means that we can quickly retrain an AI model after there is a new cyber-security hack or a new fraud situation,” Hunter told Digital Trends. “Waiting for days or weeks to retrain the model is not practical — so being able to train accurately and within hours makes a big difference.”

These massive improvements in terms of speed, combined with advances in terms of accuracy make IBM’s distributed deep-learning software a major boon for anyone working with this technology. A technical preview of the API is available now as part of the company’s PowerAI enterprise deep-learning software.

Brad Jones
Former Digital Trends Contributor
Brad is an English-born writer currently splitting his time between Edinburgh and Pennsylvania. You can find him on Twitter…
The Windows 11 24H2 update is causing even more problems
Windows 11 logo on a laptop.

The Windows 11 24H2 update had already been giving users a real headache with problems such as bugs for visual layouts and flaws for certain wallpaper apps. And now, as Microsoft confirms in a support document, some people without administrative privileges can't change the time zone in the Date & Time view, among myriad other issues related to the important Windows 11 update.

A Feedback Hub post also reports a time issue after exiting Sleep Mode, specifically after about one out of every five overnight sleep cycles. There is also a report that the time is not syncing correctly following daylight saving time. Put differently, the update doesn't break the time zone, but only affects the toggle or makes it very difficult to modify it.

Read more
The 10 best monitors for 2024: tested and reviewed
OLED demo on the Asus ROG PG27AQDM.

Editor’s note: You should expect to see tons of great monitor deals on Black Friday and Cyber Monday this year. It's pretty much a guarantee that you'll find something that suits your needs at a discounted price, so long as you keep an eye out and pay close attention to what you're buying. Whether you're looking for a killer gaming monitor deal, a high-end 5K monitor deal, or even an OLED monitor deal, or something more on the budget side, we've got your covered. Make sure to check out our other Black Friday deals for even more bargains on TV, headphones, and more.

For those seeking a superior computer setup, a cutting-edge monitor is non-negotiable. As we move into 2024, the monitor market offers a wide range of options tailored to various needs, from immersive gaming displays to high-resolution panels for creative professionals. Whether you're looking for top-tier gaming performance with fast refresh rates, crisp visuals for productivity, or a versatile all-rounder, this year’s monitors bring cutting-edge features like OLED panels, high refresh rates, and enhanced connectivity. In this guide, we'll explore the best monitors you can buy in 2024, ensuring you find the perfect fit for your setup.

Read more
ChatGPT just improved its creative writing chops
a phone displaying the ChatGPT homepage on a beige bbackground.

One of the great strengths of ChatGPT is its ability to aid in creative writing. ChatGPT's latest large language model, GPT-4o, has received a bit of a performance boost, OpenAI announced Wednesday. Users can reportedly expect "more natural, engaging, and tailored writing to improve relevance & readability" moving forward.

https://twitter.com/OpenAI/status/1859296125947347164

Read more