MATLAB and Simulink Seminars

Deep Learning with MATLAB


Deep learning applications have rapidly evolved over the past decade and are now being used in fields varying from autonomous systems to medical image processing. This session focuses on deep learning techniques to help solve problems such as image classification. We will demonstrate how you can train a deep network and then use it, and we will examine ways to better understand how a deep network works. We will also show how a new tool, GPU Coder, makes it easy to deploy deep learning algorithms to desktop and embedded GPUs.


  • Deep Learning: training and inference on CPUs or GPUs made accessible via a few lines of MATLAB code.
  • GPU Coder: Deployment of deep learning algorithms and vision algorithms onto embedded GPUs via automatic CUDA code generation.

Who Should Attend

  • Everybody interested in Data Analytics, Deep Learning, and Machine Learning.
  • Engineers targeting embedded GPU systems.

About the Presenter

Daniel Aronsson is a senior application engineer at MathWorks with specialization in signal processing and communications system design. He has several years of experience in these fields and he also works regularly with computer vision and FPGA system implementation.


Time Title




Deep Learning fundamentals

This session will cover deep learning techniques to help solve problems such as object detection, object recognition, and classification. We look at how to set up and train a Convolutional Neural Network (CNN), as well as modifying an existing CNN to fit a new purpose (transfer learning).




Automatic Deep Learning and CUDA code generation

You got your trained network – what do you do? Staying in our MATLAB-centric workflow, we show how our MATLAB code can be automatically converted to CUDA code, and then deployed on to GPUs whether on your desktop, a cluster, or on embedded Tegra platforms, including Jetson TX2.



Registration closed