Adam & Eva Inc. Breaks AI Training Speed Record With New Fully Parallel Version of its AI Architecture

SAN DIEGO, April 11, 2024 /PRNewswire/ -- The latest version of Adam & Eva's neuro-symbolic architecture AI, the world's first traceable Generative AI, achieved 3Gb per hour speeds in LLM training.

Unlike existing LLM-based transformers, Adam & Eva's neuro-symbolic AI supports a linear learning rate depending on the size of the dataset. This provides several orders of magnitude reduction in computing power, time, and costs.

"We are happy to confirm this break-through performance when training EVA, our game changing LLM." – said Paul Kewene-Hite, Adam & Eva Inc. CEO & Co-Founder – "In addition to amazing computational speeds, other remarkable features of EVA include an ability to scale that is limited only by data sets and CPUs, an option for clients to train our LLM EVA themselves, as well as meaningful savings in both CapEx and OpEx."

The 3Gb per hour results demonstrated by the latest fully parallel version of Adam & Eva's AI were performed on a 56-core CPU server hosted at Hyve Managed Hosting. Adam & Eva plans to present their LLM "EVA" during MadHat event in San Francisco on April 19 and to introduce it through public cloud providers in 2024.

About Adam & Eva Inc.
Adam & Eva, Inc. is a Delaware corporation, established in 2023 with offices in San Diego (CA), Boston (MA), Palo Alto (CA), and Haifa, Israel (www.adam-eva.ai). Adam & Eva, Inc. develops a traceable and scalable neuro-symbolic Gen AI architecture for public utilization. Unlike common transformers, Adam & Eva's AI demonstrates linear, rather than quadratic, computational complexity for both training and operational modes.

For media inquiries or more information, please contact:
Adam & Eva Inc.
Vadim Asadov, Co-Founder
858 405 9128
[email protected]

SOURCE Adam & Eva Inc.