Santa Clara, CA, USA, September 17, 2018 – Dr. Ren Wu, founder and CEO of NovuMind Inc., a leading innovator in full-stack Artificial Intelligence technologies, participated in this month’s IEEE System on Chip Conference (SOCC) in Arlington VA. With the theme of “Thinking SOCs: AI and ML,” the conference attracted AI researchers, engineers and industry experts from around the world.
Dr. Wu presented on the topic of “Leveraging the Power of Deep Learning in Intelligent Devices.” His presentation covered the potential of embedding AI in devices at the edge of the network, effectively creating an “Intelligent Internet of Things.” Dr. Wu also discussed the challenges of widespread deployment of AI at the network edge, including power consumption and processing power requirements. Until now, most AI functionality has resided in data centers or the Cloud, where expensive and highly specialized AI processing resources can be shared. However, as low-power and high-performance SOC solutions for AI become widespread, AI can be deployed much more widely, including deployment in embedded applications that require real-time, low-latency performance.
“Artificial Intelligence is more capable than it ever was before, and we are just at the beginning of what many people are calling a Fourth Industrial Revolution fueled by AI,” according to Dr. Wu. “But to realize the full potential of this requires specialized AI processing chips and modules, much better suited to Deep Learning models than general purpose chips like Nvidia GPUs.”
Dr. Wu presented details of the NovuTensor chip, with its domain-specific architecture purpose-built for deep neural network processing. In this architecture, 3-dimensional tensor computations are performed natively, resulting in a very high performance to power consumption ratio for deep neural network processing. NovuTensor is the only AI chip that achieves server-class performance while still within the power budget of embedded systems.
NovuMind helps companies put the power of AI into their products. The NovuTensor chip, along with a full stack of AI solutions support, enables OEMs to embed deep learning inference capability, even in applications where power or network access are limited. Led by AI industry pioneer Ren Wu, the NovuMind team has expertise in machine learning, chip design, and high-performance computing. The company’s headquarters and R&D are located in Santa Clara CA, with additional offices in Beijing and Guangzhou, China. For more information, please visit us at: www.novumind.com