HPE releases ML development system to help companies deploy AI at scale

We’re excited to bring Transform 2022 back to life on July 19th and virtually July 20-28. Join AI and data leaders for sensible conversations and exciting networking opportunities. Register today!


It’s a puzzle across the enterprise sector: Artificial Intelligence (AI) and Machine Learning (ML) modeling provide excellent business value in a variety of use cases. But to achieve this requires significant time and financial investment in AI infrastructure.

And many organizations aren’t there yet – meaning engineers spend most of their time doing manual tasks and infrastructure management instead of modeling, training and organizing.

Justin Hotard, executive vice president and general manager of HPC, said, “Businesses want to include AI and ML to differentiate their products and services, but often face the complexity of setting up the necessary infrastructure to create and train specific AI models.” ” And AI at Hewlett-Packard Enterprise (HPE).

Problems with AI and ML deployments throughout the enterprise

There is no doubt that investment in AI / ML is growing steadily and at a significant pace: According to Tortoise Intelligence, global investment has grown 115% since 2020, the highest year-on-year growth in two decades. Similarly, Fortune Business Insights estimates that the ML market size will increase from approximately $ 21.2 billion in 2022 to $ 209.91 billion in 2029, with a compounded annual growth rate of about 40%.

But while organizations prefer AI / ML over other IT initiatives, they lag behind in post-deployment operational issues, deployment, and often move on to different structural complexities.

In a recent survey by Comet, 68% of people reported scraping anywhere from 40% to 80% of their AI / ML experiments. This was largely due to the “sadly inadequate” budget and the mismanagement of the data science lifecycle outside the normal repetitive processes of breakdown and experimentation.

HPE for rescue

As a tool to help make this process easier and faster, HPE today unveiled a new machine learning development system. The ready-to-use system allows users to instantly create and train AI models on a scale and realize quick value. It is based on the acquisition of HPE in the summer 2021 of AI. The San Francisco startup has created an open-source AI training platform that has now transitioned to the HPE machine learning development environment.

“Users can speed up the typical time-to-value for weeks and months, from days to days, to start building machine models and realizing the results of training,” Hottard said.

Traditionally, he pointed out, a complex, multi-step process is needed to adapt infrastructure to support model development and training on a scale. This includes the purchase, setup and operation of highly parallel software ecosystems and infrastructure.

On the contrary, he said, the HPE machine learning development system is fully integrated and ready for use, with a combination of software and specialized computing, including accelerators, networking and services. It can scale AI model training with minimal code rewrites or infrastructure changes and helps improve model accuracy with distributed training, automated hyperparameter optimization and neural architecture detection – all of these ML algorithms.

The system delivers optimized computers, accelerated computers, and interconnect, supporting scale modeling for a combination of workloads. Its small configuration starts with 32 GPUs, which show about 90% scaling efficiency for workloads, including Computer Vision and Natural Language Processing (NLP), Hotard said.

For example, German AI startup Elf Alpha implemented a new HPE system to train multimodal AI, including large natural language processing (NLP) and computer vision models. The company was able to set up a new system integrating and monitoring hundreds of GPUs in just two days and began training on it in two days.

The company set up customized hyperparameter optimization and experiment tracking for collaboration, Hotard explained. AI assistants are able to search for highly specialized information in complex texts, high-level comprehension summaries, and hundreds of documents. They are also able to take advantage of specialized knowledge in the context of communication.

“By combining image and text processing in five languages ​​with an almost humane contextual understanding, the models push the boundaries of modern AI for all types of language and image-based transformative use cases,” Hottard said.

All told, the machine learning development system could improve ML team collaboration by providing a faster route to more accurate models, Hotard said, also enabling flexibility that could help future-proof AI infrastructure. It combines our proven end-to-end HPC solutions for deep learning into a system with our innovative machine learning software platform to provide a performance out-of-the-box solution to accelerate value and results with AI, ”he said. . .

Venturebeat’s mission Transformative Enterprise is about to become a digital town square for technology decision makers to gain knowledge about technology and transactions. Learn more about membership.

Similar Posts

Leave a Reply

Your email address will not be published.