On its own front, Microsoft has been keen to
get more folks developing AI, enough to recently open-source the deep learning tools it used to
build Skype Translate and Cortana for users to train their own AI. It's also
opened its arms to other tech companies, partnering with Google, Amazon, IBM
and Facebook in a coalition to
trumpet the benefits of AI and agree on best practices. The initiative left out
the Elon Musk and Peter Thiel-backed OpenAI, a research project dedicated to democratizing
artificial intelligence. But today, the non-profit announced a separate team-up with Microsoft to
run large-scale experiments on the software giant's Azure cloud platform.
Azure's
open-sourced tools appealed to the non-profit, as well as its
computation-boosting Batch and machine learning capabilities, Microsoft said in
a blog post. The partnership
also nets OpenAI early use of the software giant's GPU-based Azure N-series
virtual machines, which will be opened for general use next month.
It's
not the first AI partnership, nor the only one to rely on cloud-based systems
to run computations, but Microsoft is devoting plenty of development to boost
Azure's appeal. While their current N-series virtual machines run on Nvidia
Tesla GPUs, those will be replaced by next-generation Pascal-based GPUs in the
near future.
Microsoft is also announcing its new Azure Bot Service,
which aims to let developers speed up development of bots within the software
giant's Bot Framework and manage them in a server less Azure environment,
according to Microsoft.
No comments:
Post a Comment