|09:30 - 10:00||Registration and networking with refreshments|
|10:00 - 10:15||Welcome from CW
Robert Driver, CEO, CW
Welcome from our host The Computer Labs
Dr Ian Wassell, University of Cambridge Computer Laboratory
A word from our headline sponsor Magna
David Paul, Director, Corporate Engineering and R&D, Magna International
|10:15 - 10:30||Scene Setting: The Landscape of AI
Phil Claridge, Mandrel Systems
Peter Whale, Peter Whale Consulting
An introduction from the AI SIG Champions to the conference day, the key topics to be covered, the structure of the day, and some of the engineering design questions to be covered.
|10:30 - 10:55||The near-term impact of AI
Professor Steve Young, University of Cambridge
A review of the major algorithmic approaches and technological advances that are driving the current uptake of AI.
|10:55 - 11:20||Key and emerging AI technologies
Under the covers of Deep Learning
Theophane Weber, Google DeepMind
What is Deep Learning? How is it different from classic neural nets? How is this informed from our understanding of the human brain? What is Deep Learning is good for and not so good for.
|11:20 - 11:50||Refreshments and networking|
|11:50 - 12:15||Key and emerging AI technologies cont.
Under the covers of a range of other AI technologies
Professor Carl Edward Rasmussen, University of Cambridge / PROWLER.io
This is the counter point to the session on Deep Learning, where we explain some of the other promising areas of AI such as Probabilistic models, Reinforcement Learning (RL) and Multi-agent Systems (MAS).
|12:15 - 12:35||The revolution of speech recognition technology
Dr Tony Robinson, Speechmatics
Speech recognition technology is revolutionising the industry – but how do you make speech recognition work for you?
|12:35 - 12:55||Panel Session with audience Q&A|
|12:55 - 13:00||A word from our sponsor PROWLER.io
Vishal Chatrath, PROWLER.io
|13:00 - 13:55||Lunch and networking|
|13:55 - 14:20||Tools and Frameworks and AI Systems Engineering
Where next for AI?
Professor Neil Lawrence, University of Sheffield
A brief look at some of the most promising areas of AI research which could be brought to market in the next 2-3 years.
|14:20 - 14:40||Survey of Tools and Frameworks
Dan Neil, Benevolent AI
This talk will look at the capabilities of the increasing number of tools and frameworks for AI, and how far they can take us currently. We will then reflect on Benevolent's experience of area where we have needed to augment these with our own in-house tools.
|14:40 - 15:00||NVIDIA’s Artificial Intelligence Platform
Alison B Lowndes, AI DevRel, EMEA, NVIDIA
This talk will combine knowledge of world-wide state-of-the-art research, with NVIDIA’s ecosystem of software, research, training, support and of course hardware. Chips are one part of enabling AI, a field moving faster than chips can be produced. That momentum drives efficiency and forces agility, allowing us to enable AI across the world’s data centres, clouds & ‘at the edge’. Discussion will include the tools we opensource and optimize with partners across academia & Enterprise as well as hints along the path to AGI through neuroscience, and what that means for the real world.
|15:00 - 15:20||Why do we need another processor (solution?) for AI
Nigel Toon, Graphcore Ltd.
This talk will cover how to design a processor for Machine Intelligence. It will describe the underlying compute workload in today's and future Machine Intelligence applications. It will show how current CPU and GPU processors are limited in their ability to support this new workload and how a new type of intelligence processing unit can be developed which is much more efficient for this fundamental new era of computers.
|15:20 - 15:40||Processors for real-world AI
Dr Peter Baldwin, Myrtle Software
A number of different processor architectures have been proposed for deployment in AI applications. In this talk, we discuss trade-offs of the alternative approaches for current and future workloads.
|15:40 - 16:00||AI: open for all?
Sobia Hamid, Cambridge Data Insights
AI will have an increasing impact on business and wider society but can seem like it is driven by a very select set of people. This panel discussion explores the issues we need to resolve to have AI be open to all.
|16:00 - 16:30||Refreshment break and networking|
|16:30 - 16:50||Tools and Frameworks and AI Systems Engineering
Community-driven AI/SW/HW co-design and optimisation
Anton Lokhmotov, dividiti
Keeping up with the fast pace of AI innovation calls for an agile system approach that engages the community in a virtuous co-design and optimisation cycle, where the design of AI applications is informed by the capabilities of computer systems and the design of computer systems is informed by AI applications.
|16:50 - 17:10||System Architectures for AI
Jem Davies, Arm
AI and Machine Learning are currently generating a huge number of headlines. As the near ubiquitous computing platform on devices, Arm has a unique view on the technology and the different implementation approaches needed to make it a success. As AI/ML workloads increase in number and complexity, Jem will discuss how Arm views the challenges, options and opportunities presented, and what is being done to address these new workloads.
|17:10 - 17:30||AI on the Edge
Cyrus Vahid, Amazon (AWS)
What drives the partitioning of edge v. cloud? How intelligent can the edge be and how intelligent do we want it to be? Does the edge device really learn?
|17:30 - 17:55||Panel Session with audience Q&A|
|17:55 - 18:00||Closing remarks|