Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore why efficiency, not raw computational power, defines the future of edge AI in this 19-minute conference talk. Discover how classic compute architectures fail for low-power AI applications and learn about the new generation of neural accelerators that are revolutionizing edge intelligence capabilities. Examine real-world examples of running YOLOv8 computer vision models at just a few hundred milliwatts and understand the architectural shift toward in-memory computing. Delve into why dedicated AI accelerators significantly outperform CPUs, GPUs, and FPGAs in power efficiency, and investigate the limitations of traditional memory architectures. Learn about hybrid analog and digital architectures that enable next-generation edge devices, and understand the evolution from general-purpose models to domain-specific and context-aware AI systems. Gain insights into the future of physical AI, robotics, and intelligent systems, with practical guidance for building edge AI systems for vision, robotics, autonomy, and embedded intelligence applications.
