Neuroscience and AI: What artificial intelligence teaches us about the brain (and vice versa) | Surya Ganguli

This podcast interviews Surya Ganguli, a professor working at the intersection of neuroscience and AI, to explore similarities and differences between artificial and biological intelligence. The discussion covers the energy inefficiency of current AI models compared to the human brain (AI uses six to seven orders of magnitude more power), and explores how AI's data efficiency could be improved by adopting principles from neuroscience, such as active learning and curated data selection. Ganguli also highlights his work using AI to model and understand the brain, specifically the retina, leading to new testable predictions about its function. Finally, the conversation touches upon the potential for future brain-computer interfaces and bio-hybrid AI systems.