Evolving AI Systems Gracefully with Stefano Soatto - #502
Published:Jul 19, 2021 20:05
•1 min read
•Practical AI
Analysis
This article summarizes a podcast episode of "Practical AI" featuring Stefano Soatto, VP of AI applications science at AWS and a UCLA professor. The core topic is Soatto's research on "Graceful AI," which explores how to enable trained AI systems to evolve smoothly. The discussion covers the motivations behind this research, the potential downsides of frequent retraining of machine learning models in production, and specific research areas like error rate clustering and model architecture considerations for compression. The article highlights the importance of this research in addressing the challenges of maintaining and updating AI models effectively.
Key Takeaways
- •The research focuses on making AI systems evolve gracefully.
- •The article discusses the potential problems of constantly retraining ML models.
- •The research explores error rate clustering and model architecture for compression.
Reference
“Our conversation with Stefano centers on recent research of his called Graceful AI, which focuses on how to make trained systems evolve gracefully.”