for running neural networks in data centers and for Google TensorFlow. Google says its TPU has far outperformed both CPUs and GPUs in studies.
Achieving high-powered, energy-efficient, and ultralow-latency performance will be critical as demand for real-time AI applications grows. While devices like collaborative robots can benefit from training-based AI, more dynamic systems like autonomous vehicles, virtual assistants, and smart security and defense systems need to be able to act on the fly and may also need to be easily adjustable. Microsoft currently plans to roll out Project Brainwave via its Azure cloud services, which may end up being a hinderance depending on how readily companies are wiling to jump onto Azure. However, as new solutions continue to emerge on all sides Microsoft's FGPA approach may just be novel and powerful enough to sway potential partners.
Artificial Intelligence: What Will the Future Be?
Intelligent systems and robots will one day help us with routine tasks, handle dangerous jobs, and keep us company. But they could also make decisions that violate our ethical principles, take control of our lives, and disrupt society. Join Maria Gini — accompanied by her AI-enabled humanoid robot — during her keynote address at ESC Minneapolis Nov. 8-9, 2017, and explore state-of-the-art intelligent systems and discuss future developments and challenges. Click here to register today!
Chris Wiltz is a senior editor at Design News covering emerging technologies including VR/AR, AI, and robotics.