Deep Learning with GPUs

This is a report on the use of deep learning (DL) with neural networks powered by graphical processing units (GPU). The report is aimed at students and researchers to assist them in getting started within this vital field. This work began at the University of Leeds and led me to my current career within NVIDIA, the inventor of the GPU. These co-processors are, along with massive datasets and efficient algorithms, powering the DL renaissance worldwide. The pace of progress is so great within deep learning that NVIDIA are constantly reshaping their ecosystem and recently became the No.1 on the Green500 with their 5 Petaflop SATURN V facility, housed in Santa Clara, California. The first task it will be given, externally, is to help on the Cancer Moonshot project in the US, realising my dream that AI be used within Healthcare. Google DeepMind are also now working directly with the NHS. I hope you find this useful.

Autorentext

Unlike most, I finally managed to complete my degree after my two children grew up. This meant I approached the subject of Artificial Intelligence with a real yearning to learn. I owe Thanks to my friends, and to my big Brother, who helped me through the tough financial challenge. I dedicate this work to my awesome kids and to my Mother, Pauline.