Earth on a Chip - AI-based Autoregressive Models of our Coupled Climate System

Explore how artificial intelligence is revolutionizing climate system modeling through autoregressive approaches in this 56-minute conference talk. Learn about cutting-edge AI-based models that simulate Earth's coupled climate system, examining how machine learning techniques can create efficient alternatives to traditional physics-based Earth System Models. Discover the development of "Earth on a Chip" concepts that leverage autoregressive neural networks to capture complex climate dynamics and interactions between different Earth system components. Understand the computational advantages and accuracy improvements these AI models offer compared to conventional climate modeling approaches, including their potential for weather forecasting and climate prediction applications. Gain insights into how deep learning architectures can model the intricate relationships within Earth's climate system while maintaining computational efficiency for practical implementation.