Microsoft has stepped into delivering real-time AI with its all-new “Project Brainwave” which is a deep learning acceleration platform. It puts to use Intel’s Stratix 10 field programmable gate array (FPGA) hardware accelerator in running AI through cloud services. The FPGA chips are more flexible in comparison with traditional servers and PCs which will enable their reprogramming to undertake different tasks instead of being swapped out for new hardware. Microsoft intends to employ Deep Neural Network (DNN) into these chips to bring enhanced level of human-like thought processing to computers.
According to Microsoft engineer Doug Burger, “Running on Stratix 10, Project Brainwave thus achieves unprecedented levels of demonstrated real-time AI performance on extremely challenging models.”
This system also promises to transmit data with ultra-low latency thereby delivering results proportional to the pace of data which is greatly beneficial to applications like digital assistants and other real-time applications. It also supports deep learning frameworks including its own Cognitive Toolkit and Google’s Tensorflow with further plans of introducing it to Azure customers.
With Microsoft and Google introducing a plethora of applications in security, financial services and other industries, this platform is intended to make real-time AI processing a reality. Project Brainwave system is all set enter the market to meet industry-leading capabilities for real-time AI.