Artificial intelligence – The search for productivity

Share this post:

As organizations and individuals develop they face the same challenges again and again: Time is limited – how to get the most out of existing or shrinking resources? The same realities present themselves with the development of Artificial Intelligence (AI), and even though the journey is at its beginning, significant results are already being achieved.

The status of AI-development and AI-development platforms

Software development and tools have evolved significantly from the early 90s. While there is still a degree of manual labor in software development, automation, better language expression and visual development platforms have raised software development productivity to a completely new level. The same leap can be seen in server infrastructure development and management – with the right development environment and tools it is possible to optimize resource efficiency.

The development of AI is still very unstructured. The everyday life of a data scientist consists of a variety of different tasks, with some of these tasks being secondary but necessary in the absence of other solutions, to achieve results. For example, to be able to educate the AI, data needs to be “cleaned” by the developer, but they are often also required to manage development and processing environments. Graphical Process Unit (GPU) environments used for processing can be challenging to manage and IT-departments often lack experience with them. Open source AI-development environments are versatile; however, they often require a lot of fine-tuning. Additionally, the limited amount of memory and data-transfer speeds usually require some kind of workaround.

Efficient development platforms, environments and more

AI-development tools are being commercialized slowly, however by investing in development platforms and environment software, data scientists can focus more thoroughly on AI development, rather than spending time in developing their own processing servers or development environments. That being said, the options are still quite limited and meaningful development requires a platform and tools that fulfill the developers needs.

IBMs PowerAI does just that. It is one of the most versatile AI-development platforms available today, containing an interesting stack of solutions to enable deep learning-based AI development. It improves productivity in three significant sectors:

PowerAI contains the machine resources and software needed for efficient AI-development. The system enables the usage of familiar deep learning -environments using the latest NVIDIA V100 GPUs, that offers state-of-the-art performance for deep learning development. AI-developers can employ widely-used applications that are pre-installed and optimized for the hardware, like: Tensorflow, Keras, Theano, Caffe, Chainer, Torch, etc.

Productivity can also be enhanced with solutions that enable data scientists to do less manual work: PowerAI DL enables automation helping in optimizing AI-development, and PowerAI Vision offers a full workflow for users who are taking their first steps in deep learning development, or for users who just want to get straight down to development.

PowerAI is based on IBMs latest POWER9 architecture, which enables state-of-the-art hardware resource usage for demanding deep learning – AI-development. Large Model Support also offers an option to use RAM to support the GPUs limited processing power in AI education. The PowerAI-system has 500-2000 GB of RAM and a single GPU has 8-32 GB, this rapid and significant increase of memory in educating the AI enables more refined AI-development.

PowerAI also offers hardware- and software support for distributing the heavy neural network calculation process to multiple devices, this streamlines the building of server clusters. Without a ready-made solution such as this, developers and IT teams have to come up with their own time consuming and expensive solutions. Distributing calculation processes enables scalability and faster AI-development. Fast calculation processes and easy resource allocation for developers also enables better efficiency and reduces the unnecessary wait for results.

The fast-lane to AI-development

AI is already playing a key role in the development of society and technology and deep learning is a significant part of it. As a part of the digitalization toolset, AI automates product and service processes, enables better user experience and even helps to streamline internal processes of companies. It is critical to be part of this evolution, to gain a better understanding of individual or corporate behavior, to improve informed decision making. Companies that have invested in AI-development are already reaping the first rewards.

Quick setup IBM solutions, like PowerAI, support data scientists in their deep learning projects and can be used as part of cloud or locally behind a firewall, when handling sensitive data.

To find out more visit:

Deep learning platform

Hitting the Wall with Server Infrastructure for Artificial Intelligence

Sales Leader Power Systems, Nordic

More AI stories

NAVTICON: Making better decisions with the data at hand

NAVTICON, alongside IBM and IBM Business Partner Danicon ApS, developed an AI-based email scanning solution that delivers greater visibility into the vital business intelligence hidden in unstructured notices regarding cargo and vessel position. And with this information, users can more efficiently fill and route vessels across the globe. Business challenge As shipping companies struggled to […]

Continue reading

IBM IoT | Client Stories | Inwido: Building Smart Homes in Europe

Learn how IBM partnered with Europe’s largest window manufacturer to make homes in Europe smarter and safer with IBM IoT. Inwido is Europe’s largest supplier of windows and a leading door supplier. The company has operations in Denmark, Finland, Norway, Sweden, Estonia, Germany, Lithuania, Ireland, Poland, the UK and Austria and also exports to a […]

Continue reading

How IBM Watson AI supports Borgun’s international growth

”We can now go beyond the minimum requirements of our regulators and implement internal best practices for fraud monitoring, which will shrink our risk exposure further still. As our markets become more competitive, having a platform such as IBM Safer Payments is crucial to react quickly to meet the new challenges.” – Steen Henriksen, Chief […]

Continue reading