Scaling Zenseact's Stack Towards Autonomy on a Next-Generation DL Supercomputer

Overview

Explore how Zenseact leverages deep learning supercomputers to advance autonomous driving, using pseudo-annotations and self-supervised learning to reduce manual annotation while ensuring safety through neural simulation technologies.

Syllabus

• Introduction to Zenseact's Autonomous Driving Stack
Overview of Zenseact
Key components of the autonomous driving stack
• Next-Generation Deep Learning Supercomputers
Architecture and capabilities
Role in scaling AI models
• Pseudo-Annotations in Autonomous Driving
Definition and use cases
Techniques for generating pseudo-annotations
Reducing reliance on manual annotation
• Self-Supervised Learning for Autonomous Driving
Principles of self-supervised learning
Implementations in the Zenseact stack
Advantages in reducing data labeling efforts
• Safety Assurance through Neural Simulation
Neural simulation technologies overview
Simulating driving scenarios for safety testing
Integration with deep learning models
• Scaling and Deployment Strategies
Scaling deep learning models in autonomous vehicles
Deployment on supercomputing architecture
• Case Studies and Real-World Applications
Successful implementations at Zenseact
Case studies of autonomous vehicles using the stack
• Future Directions and Innovations
Emerging technologies in autonomous driving
Future improvements to the Zenseact stack
• Conclusion and Key Takeaways
Summary of key concepts
Impact of next-generation supercomputers on autonomous driving
• Additional Resources
Recommended reading and research papers
Online tools and communities for further learning

Subjects

Computer Science