AI lounge: Speeding up Deep Learning with Emerging Hardware
Welcome to this new AI Lounge. We are happy to present two new interesting presentations.
Poster about the AI lounge (pdf).
Coffee, tea and cookies will be served.
Looking forward to seeing you!
Huawei’s activities to deliver the next generation AI hardware
The presentation topic is AI development and Migration.
by Dr. Sun (Huawei)
Dr. Sun achieved his Doctor Degree in Researching of Video compression and has been working in Huawei for 6 years. His expertise is in Video/image processing and Cloud Computing.
Machine Intelligence at Supercomputer Scale
The talk will introduce Graphcore. It will then introduce our processor, the Intelligence Processing Unit (IPU), our IPU scale-out machinery (IPU Pods), and our software tool-chain. The tool-chain allows for expressing and running computational graphs on the hardware via frameworks such as Tensorflow. The system as a whole has been designed from the ground up to enable the discovery of next generation machine intelligence.
by Dr. Arjun Chandra (Graphcore)
Dr. Arjun Chandra received a doctorate in Computer Science from the University of Birmingham, UK in 2011. He recently joined the scale-out systems unit of Graphcore, inventors of the Intelligence Processing Unit (IPU), as a Senior AI Engineer. He currently conducts R&D activities to support and discover machine intelligence workloads and methods that scale efficiently with compute. He routinely advises Masters projects via the Norwegian Open AI Lab. Prior to Graphcore, Arjun was a Senior Research Scientist at Telenor Research, conducting fundamental and applied research on deep reinforcement learning. Before that, he was leading the AI effort at an education technology startup (studix) based in StartupLab (Oslo), and has also conducted academic research at the Universities of Manchester, Birmingham, and Oslo, on projects focussed on engineering machine intelligence funded by the UK research council (EPSRC) and the EU.