Speaker "Victor Jakubiuk" Details Back



Debunking the myth of 100x GPU vs. CPU: efficient training and inference of neural networks on commodity CPUs and embedded devices.


This talk attempts to debunk the GPU vs. CPU myth in deep neural networks applications, with examples from practice in neuroscience and autonomous navigation. The talk will discuss how to performance engineer computer vision convolutional neural networks (such as ResNet) with Caffe to achieve inference throughput on the order of 1TB/hr on a multi-core Intel CPU, without the need for the GPU. The talk will also touch on other, related, aspects of CPU engineering of a high-performance visual processing pipeline, both in the data center as well as in embedded devices.


Victor Jakubiuk is a scientist and an entrepreneur. Victor worked as a research scientist in a computational neuroscience lab at MIT CSAIL. Currently he's the Chief Scientist of - a Palo Alto-based startup providing the fastest CPU-based computational engine for neural networks. Victor holds a B.S. and M.S. from MIT and lives in San Francisco. In his spare time he loves triathlon and can be often spotted swimming in Aquatic Park.