HP Enterprise acquires AI training platform to accelerate ML modeling
Determined AI is a San Francisco-based startup that provides a software stack designed to train AI models faster, at any scale, using its open source machine learning (ML) platform. HPE says it will combine Determined AI’s unique software solution with its world-leading AI and high-performance computing (HPC) offerings to enable ML engineers to easily implement and train machine learning models to provide faster and more accurate insights from their data in almost every industry.
“As we enter the Age of Insight, our customers recognize the need to add machine learning to deliver better and faster answers from their data,” says Justin Hotard, senior vice president and general manager, HPC and Mission Critical Solutions (MCS), HPE. “AI-powered technologies will play an increasingly critical role in turning data into readily available, actionable information to fuel this new era. Determined AI’s unique open source platform allows ML engineers to build models faster and deliver business value sooner without having to worry about the underlying infrastructure.”
Building and training optimized machine learning models at scale is considered the most demanding and critical stage of ML development, and doing it well increasingly requires researchers and scientists to face many challenges frequently found in HPC. These include properly setting up and managing a highly parallel software ecosystem and infrastructure spanning specialized compute, storage, fabric and accelerators. Additionally, users need to program, schedule and train their models efficiently to maximize the utilization of the highly specialized infrastructure they have set up, creating complexity and slowing down productivity.
Determined AI’s open source machine learning training platform is designed to close this gap by helping researchers and scientists focus on innovation and accelerate their time to delivery by removing the complexity and cost associated with machine learning development. This includes making it easy to set up, configure, manage and share workstations or AI clusters that run on-premises or in the cloud.
Determined AI also is designed to make it easier and faster for users to train their models through a range of capabilities that significantly speed up training, which, says the company, in one use case related to drug discovery went from three days to three hours. These capabilities include accelerator scheduling, fault tolerance, high speed parallel and distributed training of models, advanced hyperparameter optimization and neural architecture search, reproducible collaboration and metrics tracking.
“Over the last several years, building AI applications has become extremely compute, data, and communication intensive,” says the company. “Ten years ago you could do cutting-edge computer vision research on a laptop. Today, you need a massive farm of GPUs or other specialized chips to remain competitive. These problems aren’t unique to vision or academic research – they’re affecting organizations large and small, as we are hearing from our thriving open source community.”
“To put it a different way: AI is rapidly becoming a High Performance Computing problem. HPE is already a global leader in designing and delivering High Performance Computing systems and, via their acquisitions of Cray and SGI, they have decades of experience in the space working with some of the most sophisticated users on the planet. We are thrilled about the opportunity to partner with HPE to deliver co-designed software and hardware and tackle some of society’s most pressing challenges.”
Both companies say they share the vision that driving an open standard for AI software infrastructure is the fastest way for the industry to realize the potential of AI. As a result, HPE is reportedly committed to investing in and rapidly growing the Determined Training Platform as an open source project.
In addition, says HPE, through its GreenLake cloud services for high performance computing, HPE is making HPC and AI solutions even more accessible and affordable to the commercial market with fully managed services that can run in a customer’s data center, in a colocation, or at the edge using the HPE GreenLake edge to cloud platform.
AI infrastructure ‘for everyone’ goes open source
Nvidia unveils CPU for giant-scale AI and HPC
Microsoft announces new supercomputer, large-scale AI vision
Computational ‘breakthrough’ enables AI training on mobile devices
‘Universal’ processor integrates CPU, GPU, and TPU capabilities