The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Programmatic Labeling and Data Scaling for Autonomous Commercial Aviation with Cedric Cocaud - #601
Nov 28, 2022
Cedric Cocaud, Chief Engineer of the Wayfinder Group at Acubed, shares insights into the future of autonomous commercial aviation. He discusses the challenges of data labeling and the transition to automated processes in air taxi development. Cedric highlights the importance of synthetic data in enhancing machine learning models and the complexities of integrating these innovations safely. The conversation also delves into the unique challenges of autonomous aviation compared to self-driving cars, showcasing the exciting intersection of aerospace engineering and cutting-edge AI.
AI Snips
Chapters
Transcript
Episode notes
The Elephant on the Taxiway
- An elephant was once recorded walking on airport taxiways in Africa.
- This highlights the unexpected events an autonomous system must handle.
First Autonomous A350 Landing
- Cedric Cocaud's team developed the AI for the first autonomous landing of an A350.
- The A350 is a large commercial aircraft carrying 300-350 passengers.
Data Collection Challenges
- Flying an A350 for data collection is expensive and time-consuming.
- Cedric Cocaud's team uses a smaller, more readily available Beechcraft Baron for data collection.