Developing perception algorithms for autonomous vehicles
Who is this presentation for?
- Deep learning and machine learning practitioners and IT staff responsible for deployment and operation of deep learning clusters
Developing an end-to-end hardware and software platform to power self-driving cars is no mean feat. A large part of the platform is the perception stack that enables the vehicle to perceive its environment—identifying road markings, objects around the car, pedestrians, and so on.
NVIDIA has been tackling this challenge for a number of years. Its end-to-end hardware and software platform is already widely used by hundreds of companies worldwide. When NVIDIA first started working on this program, it soon became apparent that developing perception algorithms is a tremendous challenge. Not only do you have to develop a large number of different AI modules and networks but you also need to prove that those models function safely in thousands of different conditions and locations. Progressing from the prototype that performs correctly in one specific set of conditions to a full system that operates robustly regardless of the context requires a huge amount of training, optimizing, simulating, and testing.
Adam Grzywaczewski takes a closer look at the challenges faced by those attempting to develop autonomous vehicles and explores a development methodology that addresses the complications that arise from the scale of the problem. He discusses how the knowledge gained by developing autonomous vehicles is applicable to other domains in terms of hardware infrastructure (NVIDIA POD architecture), infrastructure operation (DeepOps library), and software utilities used for model development (Clara and transfer learning toolkit), optimization (TensorRT), and deployment (TensorRT inference server).
- A basic understanding of machine learning and deep learning
What you'll learn
- Gain a deeper understanding of deep learning algorithm development and the associated challenges
- Take a deep dive into deep learning-algorithm development methodology, deep learning infrastructure, and the software required
Adam Grzywaczewski is a deep learning solution architect at NVIDIA, where his primary responsibility is to support a wide range of customers in delivery of their deep learning solutions. Adam is an applied research scientist specializing in machine learning with a background in deep learning and system architecture. Previously, he was responsible for building up the UK government’s machine-learning capabilities while at Capgemini and worked in the Jaguar Land Rover Research Centre, where he was responsible for a variety of internal and external projects and contributed to the self-learning car portfolio.
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires