10/09/2026
Complex systems, such as a power grid or even the Earth's climate, can appear stable even as they approach a sudden, potentially irreversible tipping point.
In his new book, “Reservoir Computing: Machine Learning Meets Nonlinear Dynamics,” ASU Regents Professor Ying-Cheng Lai explores how historical data can help researchers identify those hidden risks — even when the equations governing a system are unknown.
Reservoir computing combines machine learning with nonlinear dynamics to create adaptable digital twins. These computational models can simulate unseen conditions and help scientists and engineers anticipate changes before an ecosystem collapses or an engineered system fails.
In a Q&A, Lai explains how chaos differs from randomness, how digital twins can make previously invisible threats visible and why limited, incomplete or noisy data remains a challenge. He also discusses his group’s work on ecological forecasting, transformer models and quantum reservoir computing for navigation in GPS-denied environments.
ASU Regents Professor Ying-Cheng Lai explores how machine learning and nonlinear dynamics can reveal hidden risks before systems fail.