Experience AI Theory in Action at the Artificial Intelligence Conference in London

There’s no “one size fits all” when deploying AI today – instead, you’ll find a mix of applications with varying compute, storage, memory, and communications needs. Intel AI technologies are designed to support usage models as diverse as training massive, unstructured data sets to performing on-device inference. And the industries finding value in AI are equally varied. At this week’s Artificial Intelligence conference in London, jointly presented by O’Reilly Media and Intel Corporation, customers are demonstrating new uses for AI across retail, healthcare, transportation, and energy science.

These customers recognize the complexity of AI workloads and are using the diverse portfolio of Intel hardware – from CPUs to VPUs and FPGAs -to full advantage in a range of multipurpose, purpose-built, customized, and application-specific solutions. They are also leveraging Intel’s software optimizations to extract top performance, whether in the cloud or at the edge. Examples of Intel’s AI technology in action include Intel® Xeon® Processors that train high-content images in minutes for better healthcare imaging, Movidius™ Myriad™ X VPUs that detect endangered species in real-time on remote cameras in the wild, and Intel® FPGAs that provide the fastest search results Bing* has ever seen.
Visit Intel Booth #107 and attend our keynotes and technical sessions and those of our customers. Then talk with our experts to learn how Intel’s diverse hardware portfolio, designed for scale and optimized with popular frameworks, can put AI to work for you.


Keynote: AI in Production: The Droids You’re Looking For
The AI of science fiction is at work today. It’s transforming IoT at the edge, improving healthcare outcomes, automating factory assembly lines, reducing traffic, and much more. Jonathan Ballon, General Manager, Intel IOT Group, will show how Intel’s diverse portfolio of AI and computer vision edge technology is driving advances and creating opportunities to monitor and inspect where humans can’t – improving how we work and live.
Date: Wednesday, October 10
Time: 9:05 – 9:20 GMT
Location: King’s Suite
Speaker: Jonathan Ballon, General Manager, Intel Internet of Things Group

Keynote: Deep Learning at Scale – A Field Manual
In this talk, Jason Knight, Head of Software Products at Intel AI, will describe the state of the field for scaling training and inference across distributed systems from a practitioner’s point of view. He’ll dive deep into the tools, resources, and venues available for getting started today without having to go it alone.
Date: Thursday, October 11
Time: 9:30 – 9:40 GMT
Location: King’s Suite
Speaker: Jason Knight, Principal Engineer

Session: Scaling Deep Learning on AWS Using C5 Instances Using MXNet* TensorFlow* and BigDL* – From Edge to Cloud
Deep learning models are now the dominant AI approach for image, audio, and sequence data delivering very high accuracy. Many of the most popular frameworks, such as TensorFlow, MXNet, Caffe, and PyTorch*, harness the power of modern hardware by focusing on specific optimizations. This talk presents trade-offs and design choices applicable to deep learning models. It then demonstrates how to optimize cost, performance, and convergence for your deep learning models, using CPU spot instances in AWS.
Date: Wednesday, October 10
Time: 13:45 – 14:25 GMT
Location: Westminster Suite
Presenters: Gaurav Kaul (Amazon Web Services), Suneel Marthi (Amazon Web Services)

Session: The Evolution of AI at the Network Edge: How Silicon is Paving a Path for IoT Innovation
Moving out from the cloud to the network edge, AI innovation is evolving. Silicon and device capabilities are fueling better experiences with AI and, in turn, the demand for AI is pushing device-level innovation. This talk looks at trends, market examples, and how IoT’s use of AI is leading to fascinating growth in various applications. From Intel’s unique vantage point, it views platforms paving the way for new AI experiences and shows Intel’s latest IoT innovations.
Date: Wednesday, October 10
Time: 13:45 – 14:25 GMT
Location: Hilton Meeting Room 3-6
Speaker: Gary Brown, Director of AI Marketing, Intel IOT Group

Session: Reinforcement Learning Coach
Reinforcement learning, which has enabled machines to play Atari* and strategy games at above-human levels, is making its way into domains such as robotics and conversational experiences. While research is booming, the tools for developing and testing new algorithms or solutions can be boosted. That’s why Reinforcement Learning Coach, an open source Python* library, was built.
Date: Wednesday, October 10
Time: 16:50 – 17:20 GMT
Location: Westminster Suite
Speaker: Gal Novik, Data Science Manager

Session: Portability & Performance in Embedded Deep Learning: Can We Have Both?
Work on low-precision inference shows that large gains in energy efficiency can be achieved by training for quantization. Embedded runtime packages like TensorFlow* Lite and Caffe2go* offer portability over a number of platforms. This talk examines best-in-class design techniques and takes a look at the challenges and progress needed to close the portability-performance gap.
Date: Thursday, October 11
Time: 11:55 – 12:35 GMT
Location: Windsor Suite
Speaker: Cormac Brick, Principal Engineer

We’ve also worked with partners in the industry to optimize Intel® architecture for AI applications. You can check out the following sessions to learn how our partners are using Intel technology to put AI to work.

Session: Efficient Neural Network Training on Intel Xeon Processor-based Supercomputers
As part of the Intel® Parallel Computing Center initiative, SURFsara* and Intel collaborated to advance the state of large-scale neural network training on servers based on Intel Xeon Scalable processors. Hear insights on several best-known methods and real-world problems that can be solved with models efficiently trained at large scale.
Date: Thursday, October 11
Time: 14:35 – 15:15 GMT
Location: King’s Suite
Presenters: Ananth Sankaranarayanan, Machine Learning Engineer (Intel), Valeriu Codreanu (SURFsara), Damian Podareanu (SURFsara), Colin Healy (Dell)

Visit Intel in Booth 107–and Intel® AI Builders members across the event

You’ve heard about making AI a reality. Come see how it’s done, with the right software running on diverse AI hardware. Intel AI technology helps simplify solution development across heterogeneous workloads and complex applications. The world-changing work AI makes possible comes in many shapes and sizes. Our demos showcase how Intel enables them all.
If you have an interest in the Intel AI Builders enterprise partner program, be sure to visit a few of our members:
Arya.ai is presenting their solution to automate insurance claims, using deep learning.
Wipro is demonstrating anomaly detection in huge structures, including oil and gas pipelines, using machine learning.
You’ll also want to catch the session “Design to architecture and code using deep learning: Implications for GUI development” from another Intel AI Builder, Mphasis, on Thursday, October 11 at 16:00-16:40 GMT in Hilton Meeting Room 3-6.

Intel Booth 107 Demos

ARYA.AI – Insurance (BFSI – Banking, Financial Services, Insurance)
Deep learning to automate complex insurance claims
WIPRO – Manufacturing
Anomaly detection in pipelines and huge structures
MINIBOT/NEXCOM – Automotive, Electronics and Electrical industry, Education
Robotic automation by two robots, using AI vision to improve robotic control
IEI/QNAP – Healthcare
Medical Image Detection/Recognition (age-related macular degeneration)
RESONATE – Traffic Management for Rail Operations
The Luminate* platform using AI algorithms to predict outcomes in real time
ROLLS ROYCE MARINE – Intelligent Awareness*
A sensor fusion system using computer vision and machine learning models to provide real-time situational awareness at sea
INTEL AI ACADEMY – Four compelling demos:
Real Time Style Transfer AI Glasses: VR glasses transform the environment into a 3D work of art in real time.
DeepCare* Chatbot: A customer service agent trained with a generative deep learning model chats with visitors.
Smart Shopping Cart: A RaspheryPi* camera classifies the contents of a shopping basket in real time.
Pictures in Synthesis – A Visual Poetry Machine: Deep learning algorithms produce surrealistic visuals.

AI at night–come party with us!

Head over to Heist Bank on Wednesday night for a party co-sponsored by Intel and O’Reilly. It’s just a short walk from the Hilton Metropole along the Grand Union Canal. Your badge admits you to an evening that offers great food and drink, a live DJ, and a chance to connect with speakers, colleagues, and other attendees.
Date: Wednesday, October 10
Time: 19:00 – 21:00 GMT
Location: Heist Bank, 5 N Wharf Rd, Paddington, London W2 1LA, UK
Admission: Open to all conference attendees—must wear badge to get in

Follow @IntelAI for more

We hope to see you in London! Be sure to follow @IntelAI for all the news and info from The AI Conference. You can tag us using #intelAI and #TheAIConf with tweets from the show. For more information and blogs visit ai.intel.com.