High Performance Computing: The Power Behind AI Transformation
It wasn’t too long ago that AI was a futuristic concept, seen only in movies. Today, cognitive computing applications, such as machine learning and deep learning, are transforming industries across the globe. However, there is another set of technologies working silently behind-the-scenes to power AI applications: high performance computing (HPC).
Before AI entered the scene, HPC environments were already being used by many scientists and engineers to help with the overwhelming amount of data and analysis involved in their fields. So why is AI relying on HPC? In order to understand this relationship, start by picturing your dream car. It might look sharp from the outside, but if you lift the hood and it’s missing a powerful engine, it won’t be of much value. HPC is the engine of AI: it’s the infrastructure that enables applications to handle high-performance workloads, such as advanced analytics.
As more industries become dependent on their ability to access and analyze data, a new wave of commercial organizations are using HPC for the first time. Three industries in particular – banking, manufacturing and healthcare – are already seeing the impact of the HPC/AI relationship.
Banking on Less Risk
Financial services companies have been reluctant to adopt AI, as a PWC study found that two-thirds of organizations are limited by operations, resources or budgets. However, those that are investing in HPC-powered AI applications are seeing a huge impact. Some companies are using these applications for risk modeling to determine aggregate risk in financial portfolios. Others are using HPC-powered AI for real-time fraud detection, as millions of transactions can now be processed between disparate systems. There has also been a rise in high-frequency trading applications and regulatory compliance applications over the past few years, which require a high performance infrastructure to handle the millions of interactions taking place. In a sector that is defined by compliance, rapid decision making and risk management, HPC’s ability to handle complex data analysis will be extremely critical in the years to come.
Manufacturing, Safer Cities and Transportation
The manufacturing industry has been quicker to embrace AI technology, recognizing it as way to stay competitive and profitable. But with a rise in complex data-driven projects, such as smart cities and autonomous vehicles, HPC-powered AI is becoming the new normal. For example, the University of Stuttgart is using high performance computing simulation to map foot traffic in the city centers. This gives the cities’ leaders data to provide better services and safer public transportation. As more cities and industries look to implement these large data-driven initiatives, HPC will be a necessary part of the transformation. Smart vehicles rely heavily on HPC systems in designing tomorrow’s self-driving vehicles. Smart power grids require capturing, analyzing, and reporting power usage for millions of businesses and residences – a challenge ready-made for HPC. Now add smart buildings, and you can begin to imagine how important HPC systems are for managing the smart cities of the future - requiring the processing of vast amounts of data in fractions of a second.
Delivering a Healthy Dose of Technology
Some Americans panic at the thought of robot doctors, but the truth is that HPC-powered AI applications are much more likely to help doctors and medical professionals, rather than replace them. One consistent struggle for many patients and doctors is diagnosis and information sharing after lab work is complete. Patients receive lab results electronically or by mail, and then wait days for a follow-up call from their doctors for an explanation and next steps. HPC-powered AI applications can help with on-demand treatment plans, analyzing the lab results and providing a diagnosis before the patient even leaves the office. Additionally, HPC can assist with recommendations for precision and personalized medicine, which requires analyzing thousands of variables for every patient. Since its early days, HPC has also helped with drug research and it continues to be an asset in many pharmaceutical companies today as well.
AI is all about fast decision making based on advanced analytics, and HPC gives AI applications the power necessary to drive this innovation. As more industries require data integration and analysis, we will see High Performance Artificial Intelligence (HPAI) become the standard ingredient for enterprise transformation.
Jeff Reser is manager, global HPC products, at SUSE.