AI

Accelerate AI projects with the right infrastructure (Part two)

Share this post:

Last week, I talked about how implementing AI solutions in your organization is a lot like climbing a ladder–going it alone can be risky, but a good AI infrastructure is like having a trustworthy friend to help steady your climb.

In my view, a crucial piece of that infrastructure support is the IBM Power AI Enterprise Platform. This new technology offers a comprehensive infrastructure environment designed to address the key requirements for AI initiatives. In other words, to help you start climbing the AI Ladder with confidence and speed, mitigating risk.

Let’s take a look at how PowerAI Enterprise helps accelerate each step of your AI journey.

1. Get up and running

IBM PowerAI Enterprise is designed to substantially reduce the time to get up and running because it comes pre-loaded with the most popular deep learning frameworks, including all required dependencies and files, precompiled and ready to deploy. This matters.

2. Prepare your data

Further, IBM PowerAI Vision—which is part of PowerAI Enterprise—helps democratize data preparation. Designed for clickers, not coders, PowerAI Vision offers easy-to-use interfaces for data preparation and incorporates auto data labeling and data regularization capabilities so that people other than data scientists can help prepare data.

Quite simply, line of business (LOB) users can get a basic model built and even expose it as a RESTful API without technical skills. This means more of the organization can participate in the process, making the entire deployment more likely to succeed.

3. Build and train models

Building and training your AI models are typically some of the most time-consuming, iterative tasks in realizing the value of AI. Depending on your infrastructure, producing accurate results might take days, weeks or even months.

However, PowerAI Enterprise capitalizes on IBM Power Systems Accelerated Compute Servers to help reduce the time to build and train models. These servers are the same ones used as building blocks for the world’s top-ranked supercomputer—the Summit system at the US Department of Energy’s Oak Ridge National Laboratory. And you’re using the very same building blocks as they are in designing your AI solution.

IBM Power Systems AC (Accelerated Compute) infrastructure was designed from the ground up for AI, high-performance computing (HPC) and supercomputing workloads. Built on IBM POWER9 processors and NVIDIA GPUs—and using high-speed NVIDIA NVLink connections among CPUs and GPUs—not just GPU to GPU like traditional infrastructures—they are designed to deliver outstanding performance for model building and training.

By using these servers in conjunction with software built to scale these frameworks (IBM Distributed Deep Learning Library (DDL) and Large Model Support are examples), you can complete model building and training much faster—using fewer servers—than with x86 servers. Just think about what that could mean to your organization.

4. Fine-tune and deploy

To further improve the accuracy of models, you need to tune them. But model tuning is another complex, time-consuming task—and one that demands data science skills.

Why is tuning so challenging? In my opinion, partly because data scientists first must determine which hyperparameters should be tuned. IBM PowerAI Enterprise offers auto-tuning capabilities to help data scientists home in on the right hyperparameters to tune—the ones that will deliver the greatest impact.

I like to think of this feature like a modern-day transmission for a car. If I’m brand new and learning to drive (like a junior data scientist), I can drive my car safely and get going. If I’m an expert tuner, I can choose to drive whichever way I want.

5. Maintain accuracy

How do you maintain the accuracy of your models over time? You load in more data and go through the entire process again, from data prep through fine-tuning and deployment. It is a continuous, iterative loop. Accelerating each step in this AI journey helps you to iterate faster so you can better maintain—and improve—the accuracy of your model.

Taking the next steps

For many organizations, the AI journey can be painfully slow. But PowerAI Enterprise helps speed the time to value for AI initiatives. It provides frameworks, tools and powerful, accelerated servers that work together to speed the AI journey.

So next time you’re looking at that daunting climb on the AI Ladder, think PowerAI Enterprise and yell “Upward and onward!”

Learn more about PowerAI Enterprise.

Read part one of this series here.

Vice President, Cognitive BigData Systems

More AI stories

Introducing SpectrumAI with NVIDIA DGX

Artificial intelligence, machine learning and deep learning projects are happening in every industry and research team across the globe. According to IDC[1], by 2019 40 percent of digital transformation initiatives will use AI services; by 2021, 75 percent of commercial enterprise apps will use AI, over 90 percent of consumers will interact with customer support […]

Continue reading

Building your AI data pipeline

Artificial intelligence, the erstwhile fascination of sci-fi aficionados and the perennial holy grail of computer scientists, is now ubiquitous in the lexicon of business. Now more modern-business-imperative than fiction, the world is moving toward AI adoption fast. According to Forrester Research, AI adoption is ramping up. 63 percent[1] of business technology decision makers are implementing, […]

Continue reading

Accelerate AI projects with the right infrastructure (Part one)

If you read the tech press, you’ll hear that artificial intelligence (AI) is all the rage these days.  And whether or not they are doing it well, everyone is saying that they’re engaging in AI. One thing is certain: organizations across industries are running fast to be a part of the AI revolution. In this […]

Continue reading