Deep Learning Tutorial

Basics of Deep Learning for Vision, Robotics, and Autonomous Vehicles.

Frank Domoney
Director at Glencroft Ltd, Lavenham Suffolk, UK

By Frank Domoney, Director at Glencroft Ltd, Lavenham, Suffolk, UK

May 11th, 2018 (Afternoon – 15:00-18:00), Dawliz hotel, Rabat, Morocco

Free of charge – mandatory online Registration (fill out the below form)

Adding machine vision to vehicles and factory robots enhances their productivity enormously.
Machine Vision enabled by CUDA, Deep Learning, and GPUs is readily available for download and implementation on any Linux machine running Ubuntu 16.06 or higher with a powerful GPU card, for example GTX1080Ti Jetpack 3.2 which allows models to be trained on the PC/GPU and then downloaded to the TX2 development kit.

The TX2 can then be used for Image Recognition, Classification, Object Detection and Scene Segmentation which are the essential technologies for autonomous vehicles.

The same technologies along with image captioning can also be used in intelligent cameras to analyse scenes and provide Kolmogorov’s compressed reporting in Smart City Environments. The Metropolis project and Deep Stream extends this to analysis of video streams

The NVIDIA Deep Learning SDK provides powerful tools and libraries for designing and deploying GPU-accelerated deep learning applications. It includes libraries for deep learning primitives, inference, video analytics, linear algebra, sparse matrices, and multi-GPU communications. Car kits are available to allow integration of the embedded TX2 modules in smart cars

The same observation applies to Robotics where vision systems used with pick and place machines allow accurate and repetitive assembly and construction. A simple learning system for robotic learning can be sourced for less than GB £5000, which includes a Robotic Arm, a conveyor belt and a sliding rail.

NVIDIA ISAAC is specialized in software development and integration in complex ICT environments. The ISAAC development kit is presently accepting applications.

As this tutorial is not a fully academic exercise but it is also targeting to foster enterprise creation, Team Selection and building will be important. Candidates will be expected to demonstrate the ability to teach themselves and contribute to adding value to the project. Furthermore, a Hackathon will be held to identify potential team members.

  1. Position:

    If "Other", please specify:

  2. Family Name (*)
  3. Given Name(*)
  4. Affiliation(*)
  5. Speciality (*)
  6. City (of your affiliation)(*)
  7. Country: (*)
  8. Paper ID (State a paper ID of your submissions to ICMCS'18, if any):(*)
  9. Email (*)
  10. Phone (*)

Video announcement

Important dates

Submission deadline March 14 18, 2018
Notification to authors March 31 April 6, 2018
Final version due April 10, 2018
Registration deadline April 16, 2018
ICMCS'18 Conference May, 10-12, 2018

Previous Editions