Present Blog-IT thought leadership in Canada

PowerAI: harness the full potential of deep learning

Written by present | Jan 31, 2019 12:39:00 PM

 

What is one of the big differences between computers and humans?

Would you say that humans would learn from their experiences whereas computers would execute instructions that were explicitly programmed?

This is no longer the case with artificial or cognitive intelligence.

 

What is this so-called artificial intelligence?

You need to see it like a Russian doll. The biggest doll is artificial intelligence. You open it, and it contains machine learning. You open this second doll and it contains a third which is deep learning.

To continue our comparison with humans:

  • Artificial intelligence is like teaching a student the information you want him to learn so he can pass his exam.
  • Machine learning is comparable to providing the student with the course book so that he can learn on his own to prepare for the exam.
  • Deep learning is similar to machine learning except that the student learns from his mistakes to be able to pass the exam in the end.

 

What is the place and the impact of these technologies in our universe?

We are not always aware of it, but they are already part of our everyday lives. They are present in search engines, in our cameras for the detection of faces, in the chatbots of many web sites to obtain information about products, in medical imaging to facilitate the interpretation of x-rays or MRIs, and in real-time translation.

They are also present in assistants such as Google, Amazon (Alexa) or Microsoft (Cortana).

Companies are looking to leverage artificial intelligence to extract ideas from data to provide customers with a better experience and maximize revenue.

Fraud detection, chatbots, recommendation engines, are just a few examples of deep learning projects that companies are deploying.

Artificial intelligence is becoming widespread and, like so many other trends in the digital world, it is also fed by data, a lot of data, but also by adequate computing performance.

 

What do you need?

What software is required? What computing and storage infrastructure do we need? Should you hire more data scientists to carry out the project?

So many questions that you must answer, after first having identified the problem to solve and the relevance of the data you have or can have.

Data experts provide new and valuable information from a wide variety of relational, semi-structured, and unstructured data sources.

They do this by taking advantage of an accelerated modern computing infrastructure, on which they execute powerful machine learning and deep learning algorithms.

However, they have difficulties in optimizing their productivity and reducing the time spent on low-value tasks, given the nature, variety and disparity of the open source software used.

At the same time, there are not enough data experts to meet the demand, which is why their pay is steadily growing.

 

IBM PowerAI

PowerAI is an integrated platform of enterprise software that combines popular open source deep learning structures, efficient development tools, and accelerated IBM® Power Systems servers to meet the demands of your deep learning projects.

PowerAI runs on IBM Power8 or Power9 servers.  By optimizing connectivity between IBM processors, NVIDIA GPUs, memory, and I / O, the deployment of artificial intelligence applications is dramatically accelerated.

 

 

The PowerAI range includes:

  • PowerAI: open source software well suited to developers and computer scientists who are just beginning their development efforts and prototypes.
  • PowerAI Enterprise: A comprehensive, in-depth learning suite for organizations looking to quickly expand their AI applications.
  • PowerAI Vision: A suite of comprehensive enterprise support tools for labeling raw data sets for training, creating, and deploying vision models based on deep learning.

 

The advantages of PowerAI

IBM PowerAI offers the following benefits to data experts and more broadly to companies:

  • Integrated solution designed for production
  • Easy to use and deploy
  • Rapid deployment so customers can get started immediately
  • Optimized performance so users can create value faster
  • Ability to learn the models much faster
  • Expanded and more accurate support of the models
  • IBM support of the entire solution including open source DL frameworks.

 

Conclusion

  • IBM PowerAI is the RedHat of machine learning and deep learning.
  • In many use cases, PowerAI helps companies solve problems that have a critical time dimension. Think about the diagnosis of diseases, and the detection / prevention of fraud in real time. It's about increasing both the speed and accuracy provided by the artificial intelligence solution.
  • PowerAI makes deep learning, machine learning and artificial intelligence more accessible and more efficient. By combining this software platform for deep learning with IBM Power Systems, organizations can quickly deploy a fully optimized and supported platform for machine learning with outstanding performance.

Present is an IBM Gold and NVIDIA partner, and our consulting services can help you in your artificial intelligence projects.  Contact us at present@present.ca or 1-866-426-2748 to take the discussion further.