AI at the edge: Ultra efficient AI on low power compute platforms
4,652 views · Published 10 August 2018 · 23:59 · Indexed 29 October 2026
Channel: GeekWire · 2018 · Science & Technology
Presentation by Sophie Lebrecht, Director of Operations, xnor.ai at 2018 GeekWire Cloud Tech Summit: geekwire.com/cloudsummit Advancements in deep learning have increased opportunities to use AI in a wide number of applications. However, traditional deep learning models demand massive computing power and therefore are often tied to cloud-based data centers. The need for expensive compute platforms and internet connectivity limits the widespread use of AI in many real-world situations. This talk will explore new approaches to highly efficient deep learning and present novel ways to tightly couple hardware and software development such that state-of-the-art deep learning models can be ported into resource constrained low cost hardware platforms. This novel approach creates the potential for all devices to operate smartly and efficiently.
