RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2

Determined AI

Company

Content

Owners

American startup Determined AI, which is developing ON for accelerated learning, artificial intelligence was launched in 2017. By June 2021, the startup attracted more than $13 million in funding, GV Alphabet, LG Electronics and CRV invested in it.

History

2021: HPE bought Determined AI

In mid-June 2021, HPE announced the acquisition of a software developer to accelerate the training of artificial intelligence algorithms Determined AI. The financial terms of the transaction are not disclosed.

This acquisition will enhance HPE's machine learning capabilities at a time when AI is becoming an increasing priority for customers of the company's supercomputers and servers.

The Determined AI platform accelerates the development of machine learning projects by eliminating the need to manually configure the AI learning environment. Moreover, in the absence of fault tolerance functions, such as those of Determined AI, server failure may require engineers to restart the training cycle from scratch, which further increases the duration of the project.

HPE bought software developer for accelerated artificial intelligence training Determined AI

Another way that Determined AI can accelerate the development of AI models is to optimize the software-level learning process itself. For this, the startup platform uses a customized version of Horovod, an open-source AI training platform originally released by Uber Technologies in 2017. The startup says its version of the platform may in some cases accelerate training tenfold, reducing the duration of some projects from days to hours.

HPE will use Determined AI technology to expand its computing capabilities. To this end, the company plans to combine the startup with the high-performance computing and critical solutions division, which includes the supercomputer division.

The Determined AI platform should also be useful for customers of HPE machine learning servers. HPE has several lines of machines with a graphics processor for machine learning projects that do not require a supercomputer, but still require significant local computing power.[1]

Notes