Intel AI at The AI Conference in New York 2018

From car to cloud to edge to data center, Intel is powering AI today in organizations around the world.  Intel will demonstrate how it’s enabling organizations to take AI into their own hands at the Artificial Intelligence Conference in New York City.

Intel will offer keynotes, a tutorial, sessions and several experiences and demos in booth #200 to demonstrate how its AI portfolio offers a flexible set of solutions to meet power, performance and cost considerations. Intel will also show how it’s paving the path for the AI of the future with significant research investments in novel compute architectures that break through the limitations of traditional AI products. Intel supports a new class of purpose built, “AI by design” hardware to handle the most demanding AI workloads of the future.

Sessions

Our two keynote sessions will highlight how Intel is moving the AI ecosystem forward to accelerate the adoption of solutions, as well as looking at a few examples of AI in the real world by showcasing companies who are putting AI to work to solve critical business challenges. We’ll also be presenting a number of additional sessions covering the full Intel AI portfolio from edge-to-cloud.

Keynote: Intel AI for the Enterprise Ecosystem

The Intel AI portfolio includes hardware and software solutions that span use cases and edge-to-cloud implementations, rooted in extensive expertise in data science and research. Fiaz Mohamed, Business Development, Intel AI Products Group and Brigitte Alexander, Managing Director, Artificial Intelligence Partner Programs, Intel explain how Intel AI solves today’s business problems and how Intel’s partner ecosystem is accelerating the adoption of solutions built on Intel technology.

Speaker: Fiaz Mohamed, Business Development, Intel AI Products Group
Date: Tuesday, May 1, 2018
Time: 9:40 to 9:50 a.m.
Location: Grand Ballroom

Keynote: Increasing Business Results through AI in the Entertainment Industry

In this fireside chat, Justin Herz, executive vice president at Warner Bros. and Fiaz Mohammed will discuss how artificial intelligence can improve content discovery and monetization.  AI has assisted Warner Bros. with better predictions of ticket sales, distribution forecasting, and hyper-segmentation of their audience as well as created more efficient and compelling marketing assets.  In collaboration with Intel AI technologies, Warner Bros. is just scratching the surface of what’s possible.

Speakers: Fiaz Mohamed, Business Development, Intel AI Products Group and Justin Herz, Executive Vice President, Digital Product, Platform and Strategy, Warner Bros.
Date: Tuesday, May 1, 2018
Time: 8:50 to 9:00 a.m.
Location: Grand Ballroom

Tutorial: Accelerate Deep Neural Networks at the Edge with the Intel® Movidius™ Neural Compute Stick 

Market research estimates there will be as many as 20 billion connected devices in the market by 2020. These devices are expected to generate billions of petabytes of data traffic between cloud and edge devices. The influx of data and devices is driving demand for preprocessing data at the edge, specifically to power on vision-based devices like smart cameras, drones, robots, and augmented/virtual reality (AR/VR). Intel® Movidius™ VPU technology and AI edge solutions portfolio help developers and data scientists pioneer the low-power intelligent edge devices segment.

Intel’s Lead Developer Evangelist Ashwin Vijayakumar will give an overview and demo of Intel Movidius Neural Compute Stick, a miniature deep learning hardware development platform used to prototype, tune, and validate AI programs, specifically deep neural networks.

Speaker: Ashwin Vijayakumar, Lead Developer Evangelist, Intel
Date: Monday, April 30
Time: 9 a.m. to 12:30 p.m.
Location: Regent Parlor

Session: High-throughput Single-shot Multibox Object Detection on Edge Devices using FPGAs

Deep learning is becoming increasingly popular for visual understanding use cases on edge devices, such as image classification and object detection. Additionally, there is demand for running basic computer vision and deep learning models on edge devices due to concerns about privacy and security. There are several implementations of SSD with deep neural nets, such as GoogLeNet and VGG, which provide low-throughput object-detection solutions on edge devices.

Intel Software Engineer Srinivasa Karlapalem demonstrates a new SSD network with SqueezeNet for high-throughput single-shot multibox object detection (SSD) on edge devices using FPGAs, specifically for surveillance. This network is used for multiclass object detection on IoT platforms such as Apollo Lake and SkyLake, accelerated with Arria 10 FPGAs.

Speaker: Srinivasa Manohar Karlapalem, Software Engineer, Intel
Date: Tuesday, May 1
Time: 11:05 to 11:45 a.m.
Location: Grand Ballroom West

Session: Deploying Deep Learning on the Cloud

In this presentation, I will discuss how the cloud can be used effectively to deploy Deep Learning and what factors can be taken into account to do so cost effectively. In many situations, the cloud infrastructure’s distributed nature can be leveraged to scale Ml/DL applications using CPUs, for example. I will give examples of when and how that can be done, and the corresponding benefits, challenges, and opportunities.

Speaker: Alejandro (Alex) Jaimes, CTO & Chief Scientist @ Acesio
Date: Tuesday, May 1
Time: 11:55 a.m. to 12:35 p.m.
Location: Concourse A

Session: How Artificial Intelligence Helps Advance Day-to-day Quality and Maintenance Decisions

In manufacturing, software development, and aerospace, tech-op teams need to make critical decisions on the spot with very little information: Where and when has a similar problem occurred? How was it resolved? Who has real-world experience fixing it? What’s causing the issue, and should I expect it to happen again?

The speakers will share use cases of cognitive AI-based applications helping technical professionals make more confident decisions to solve their most pressing issues.

Speaker: Jacob Graham, Director of Product Management, Intel and Mallika Fernandes, AI Innovation lead for testing in Accenture
Date: Tuesday, May 1
Time: 1:45 to 2:25 p.m.
Location: Morgan

Session: Build Deep Learning-powered Big Data Solutions with BigDL

BigDL: Distributed Deep Learning is one of the leading deep learning libraries for Apache Spark, and is an Intel open source project. BigDL has model interoperability support with popular DL frameworks like Caffe*, Torch*, TensorFlow*, and Keras, OpenCV on Spark*, easily deployable Docker container support and a model zoo of widely used neural network algorithms. On the performance side, BigDL: Distributed Deep Learning takes advantage of the new Intel® Xeon® Scalable processors and Intel Math Kernel Library for training (node scaling) and inferencing (leveraging model quantization features). Recommendation engines, customer-merchant propensity models, large-scale image similarity search and inferencing are just some of the current BigDL use cases.

Sergey Ermolin, Solutions Architect and TPM, Deep Learning, Spark Analytics, and Big Data Technologies at Intel, will share some of the many use cases that are examples of what can be built with BigDL and will detail the latest features, and what the future holds for this technology.

Speaker: Sergey Ermolin, Solutions Architect and TPM, Deep Learning, Spark Analytics, and Big Data Technologies, Intel
Date: Tuesday, May 1Time: 4:50 to 5:30 p.m.
Location: Grand Ballroom East

Session: An End-to-end Video Analytics Solution for Surveillance and Securing High-value Assets

There is an increasing demand for running basic computer vision and deep learning models on edge devices for quick response/alarm/alert generation, privacy, security, etc. The information from the edge is then sent to an on-premises computer and then on to the cloud for further processing, such as tracing the subject and training the artificial intelligence (AI model). The challenge is to optimize and balance this end-to-end solution.

Harsh Kumar, Business Development Manager at Intel will detail the technologies to use and explore challenges system developers may face, from the number of cameras and their resolutions to the number of intelligent gateways to optimize performance, detect and correct anomalies. Kumar will also dissect the amount of data sent to the cloud or kept on-premises and show how all these challenges can be addressed by modeling and simulating end-to-end solutions using Intel® CoFluent™ for  IoT.

Speaker: Harsh Kumar, Business Development Manager, Intel
Date: Wednesday, May 2
Time: 11:55 a.m. to 12:35 p.m.
Location: Grand Ballroom East

Session: Classifying images in Spark*

Volunteers around the world increasingly act as human sensors to collect millions of data points. A team from the World Bank recently trained deep learning models to confirm that photos gathered through a crowdsourced data collection pilot matched the goods for which observations were submitted. They used Apache Spark and BigDL: Distributed Deep Learning that enables data engineers and scientists to write deep learning applications in Scala or Python as standard Spark programs, without having to explicitly manage distributed computations.

Yulia Tell, Technical Program Manager at Intel and Maurice Nsabimana, Statistician at the World Bank, will walk through how to get started with BigDL, which runs in any Apache Spark environment whether on-premises or in the cloud, and explain how to write a deep learning application that leverages Spark* to train image recognition models at scale. They will also detail a collaborative project to design and train large-scale deep learning models using crowdsourced images from around the world.

Speakers: Yulia Tell, Technical Program Manager, Intel, and Maurice Nsabimana, Statistician, World Bank Development Data Group
Date: Wednesday, May 2
Time: 2:35 to 3:15 p.m.
Location: Grand Ballroom West

Session: AI at the Edge with Intel® Moviduis™ Technology

Intel® Movidius™ technology is reducing the complexity of developing custom circuit boa rds and allowing developers and companies to prototype AI applications with the

Intel® Movidius Neural Compute Stick. Kathleen Kallot, Product Marketing Manager at Intel, and Augustin Marty, Cofounder and CEO Deepomatic, demonstrate how the newly announced “Intel AI: In Production” program makes it easier to bring these designs to market.

Speakers: Kathleen Kallot, Product Marketing Manager, Intel, and Augustin Marty, Cofounder and CEO Deepomatic
Date: Wednesday, May 2
Time: 4:50 to 5:30 p.m.
Location: Nassau East/West

Booth Activities – Intel booth #200

Visit the Intel booth to see demos from both Intel and partners on a wide range of AI use cases. We’ll also have info on the Intel® AI Academy, a resource for developer education, as well as information on talent recruiting for Intel.

Intel demos:

  • Intel® AI Academy: Essential Tools, Innovative Solutions, and Community Collaboration for AI Developers, Data Scientists, and Academics
  • Low Power Deep Learning at the Edge with Intel® Movidius™
  • Better Business Decisions with Intel® Saffron™ AI
  • Enhancing Motorsports Viewing Experience with Intel AI

Partner Demos:

  • AI in Live Automated Sports Production with Keemotion
  • Predicting Pipeline Leakage
  • Deep Learning Using GAN
  • In-Memory Real-Time Analytics Using GigaSpaces and Intel’s BigDL for Instant Impact
  • Solving Two Language Problems in AI with Julia

You can also see the following Intel partner demos in Demo Alley:

  • Understanding Natural Language at Scale with Gamalon
  • Automate Building Machine Learning Models for Your Data with DataRobot
  • How Automating Optimization Improves Machine Learning Performance
  • Streamline Data Management for Training with Mighty AI’s Ground-Truth Management Platform

Intel invites you to experience its AI portfolio through use cases, demos and sessions, and also learn how Intel is preparing the way for the future of AI –- all happening at the Artificial Intelligence Conference in New York. Make sure to follow @IntelAI for all of the news from the conference, and tag us using #intelai with Tweets from the show!