Towards Trustworthy AI - Reflections on Learning and Smale's Legacy

Explore a reflective conference talk examining two decades of machine learning research inspired by Steve Smale's foundational work in computational learning theory. Delve into the speaker's journey as Smale's final PhD student at UC Berkeley, tracing the evolution from Smale's 18th problem posed in 1998 to contemporary challenges in developing trustworthy artificial intelligence. Discover how dynamical algorithms and topology learning intersect with fundamental questions about the limitations of intelligence. Learn about the mathematical foundations underlying modern AI systems and the ongoing quest to create more reliable and trustworthy machine learning models. Gain insights into how classical mathematical problems continue to influence cutting-edge research in artificial intelligence, while honoring the legacy of one of mathematics' most influential figures on his 95th birthday.