H2O makes deal with NVIDIA to speed machine learning on GPUs


H2O.ai, supplier of an open source AI platform aimed at helping businesses deploy deep learning to solve complex problems, has announced a partnership with NVIDIA, the GPU processing company, to optimize its products on the GPU platform.

H2O’s AI products aim to enable customers to train machine learning and deep learning models up to 75x faster, compared to CPU-based solutions. Potential use cases for the GPU integration include customer service, fraud prevention, financial advising and healthcare personalization.

H2O is a founding member of the GPU Open Analytics Initiative that aims to create an open framework for data science on GPUs. As part of the initiative, H2O’s GPU edition machine learning algorithms are compatible with the GPU Data Frame, the open in-GPU-memory data frame. H2O can read the data frame and run machine learning in the GPU memory directly.

Announced in May and in addition to H2O, the Initiative has attracted BlazingDB, Graphistry, Continuum, MapD and Gunrock from UC Davis led by CUDA Fellow John Owens to contribute technical expertise.

Sri Ambati, CEO and co-founder of H2O, stated, “H2O.ai is excited to unveil our growing partnership with NVIDIA to bring interpretable, fast and accurate algorithms to GPUs. H2O.ai with NVIDIA GPU acceleration brings a high-performance, cloud-neutral learning and inference stack for enterprise AI communities.” 

In business for five years, H2O is now in use at over 9,000 companies by over 80,000 users, including users in over a third of Fortune 500 companies. 

Vinod Iyengar, director of marketing for H2O, said the open source model has worked well for H2O. “There is no way a company of our size could have software adopted by 9,000 organizations without using open source. We are slowly becoming the de facto standard for machine learning and deep learning.”

The business model is to offer support for companies using H2O  as they get into production. “Companies have many models going into production; once they do that, they need our data science and production support,” Iyengar said.

Customers include Kaiser Permanente, Capital One, Comcast, Progressive Insurance, Transamerica, Macy’s and Walgreens. 

For more information, go to www.h2o.ai.

– By John P. Desmond, AI Trends editor