4 December 2019
6:00 - Doors open. Networking. Drinks & Pizza
6:45 - Introduction
7:00 - 7:25 Talk 1: The IPU, TensorFlow and everything in between by Dave Lacey, VP of Customer Engineering at Graphcore
7:25 - 7:30 Q&A
7:30 - 7:55 Talk 2: Deep learning for classification of Attention Deficit Hyperactive Disorder (ADHD) using functional MRI by Atif Riaz, Researcher at City, University of London
7:55 - 8.00 Q&A
Speaker: Dave Lacey, VP of Customer Engineering at Graphcore
Title: The IPU, TensorFlow and everything in between
Abstract: This talk introduces the IPU - a brand new type of processor for the next generation of machine intelligence computer systems. The presentation will cover some of the design philosophy behind the IPU and how it enables you to develop breakthroughs in machine intelligence.
As well as the hardware, a good machine intelligence solution needs a quality software stack to allow people to use it. The second part of this talk will cover the Poplar software stack and how the parts of the stack build up to provide a comprehensive TensorFlow port for IPUs.
Bio: Dave Lacey is VP of Customer Engineering at Graphcore overseeing application development and customer support as people use the Graphcore IPU. Before building the customer engineering team he worked on the programming environment and software stack for the IPU. He has a PhD from the University of Oxford in Computer Science and has over 18 years of experience in research and development of programming tools and applications in many areas including machine learning, HPC and embedded systems.
Speaker: Atif Riaz, Researcher at City, University of London
Title: Deep learning for classification of Attention Deficit Hyperactive Disorder (ADHD) using functional MRI
Abstract: Brain disorders have emerged as one of the greatest threats to human health. Mental, neurological and substance use disorders constitute 13% of the global burden of disease exceeding both cancer and cardiovascular diseases. Despite the advances in imaging technologies, proper clinical diagnosis is not well established and in most cases diagnosis of a neurological disorder is achieved based on physical observations. Deep learning is the best available tool to help medical experts for the diagnosis of a brain disorder. The talk aims to explain the application of deep learning methods particularly CNNs, using the Tensorflow, in the medical domain for diagnosis of brain disorder and understand brain functions.
Bio: Atif is a researcher at the City, University of London. During his research, he has explored different machine learning and deep learning methods in the domain of medical imaging for the classification of brain disorders based on functional MRI data. His proposed novel methods have achieved better classification performance as compared to the state-of-the-art on the ADHD dataset. During his research, he has multiple journal and conference publications.