Deep Learning Neural Network Acceleration at the Edge

Explore the growing importance of edge computing in deep learning neural network acceleration through this informative conference talk. Delve into the challenges posed by increasing data from sensors and stringent latency requirements, necessitating machine learning inference at the edge. Examine the various neural network acceleration options available in recent Arm-based platforms and their potential for creating intelligent devices. Gain insights into the risks of fragmentation and effort duplication when multiple frameworks support multiple accelerators. Learn about existing neural network frameworks, model description formats, accelerator solutions, and low-cost development boards. Discover ongoing efforts to identify the best technologies for consolidation and enabling competitive innovation across vendors. Benefit from Andrea Gallo's extensive experience in the field, including his work with Linaro, STMicroelectronics, and early involvement with ARM technology.