Thursday, 30 November 2017

Slow human machine learning experts need not apply. AI is faster and better.

Supercomputer rapidly builds better neural networks (e.g. AI for science). (Slow human machine learning experts need not apply.)

Auto-generated networks can be produced quickly, in a matter of hours as opposed to the months needed using conventional [manual] methods.

It is designed to evaluate, evolve, and optimize neural networks for unique datasets. Scaled across Titan's 18,688 GPUs, MENNDL can test and train thousands of potential networks for a science problem simultaneously, eliminating poor performers and averaging high performers until an optimal network emerges. The process eliminates much of the time-intensive, trial-and-error tuning traditionally required of machine learning experts.

Next... Summit could potentially deliver exascale-level performance for deep learning problems - the equivalent of a billion billion calculations per second.

Algorithm leverages Titan supercomputer to create high-performing deep neural networks

Innovation Future Forum
» Exclusively for UK employees and free!
» Innovative ideas to solve your challenges
» Q&A about creativity and innovation
» and much more.

Modern Innovation and Creativity Training Courses
Powered by Innovation Future Specialist, the UK's most modern innovation provider.

No comments:

Post a Comment