The Latest in Artificial Intelligence Trends, Technologies, and Best Practices

Computer Vision at the edge is going through a huge transformation due to the advent of machine learning. In this talk, we highlight how Arm Partners Roviero, Deeplite, Plumerai are revolutionizing computer vision on the edge through machine learning on Arm platforms. During this talk, they will present a variety of approaches which they use to perform various computer vision tasks either running conventional neural networks, binary neural networks on hardware or software.

Since modern AI chipsets have different strategies for efficient operations, most neural network models may not be sufficiently optimized for these devices in terms of latency and memory footprint.

In this talk, we present how we make popular neural models be efficiently deployed on Ethos-U65, a newly launched Micro NPU. To this end, we first examine various operation forms (e.g., convolution types and filter size) and identify suitable operations to improve the accuracy-latency trade-off. 

Based on this investigation, we carefully redesign well-known convolutional blocks (e.g., inverted residual blocks and ghost blocks) and use these blocks to replace computationally inefficient blocks in given models. We demonstrate that the model variants obtained by our approach can significantly reduce inference time as well as memory budget without noticeable performance drops on Ethos-U65.