December 4, 2026
Location: RUPPERT (Uithof) , room DMetastability and noise-induced tipping are two related phenomena in stochastic processes, emphasising different questions. Metastability refers to the very long time a system may need to reach its equilibrium state. It has been widely studied in the context of statistical physics, and is best understood in reversible situations. Tipping can occur when a driven multi-stable system approaches a critical transition or bifurcation point, and the randomness induces a transition to a different metastable state before the actual transition point is reached. It occurs in non-reversible systems, and has important applications in climate dynamics and similar areas.
In a first part of these lectures, we will discuss sharp estimates on mean transition times in reversible metastable systems, with an emphasis on systems with many equilibrium states and involving symmetries. Examples include a discretized Allen-Cahn equation with conserved total mass, and an XY/Kuramoto model on the one-dimensional torus.
The second part of the lectures will focus on non-reversible systems and tipping. We will discuss in particular oncentration results around stable equilibria of time-dependent stochastic differential equations and SPDEs, sharp estimates on mean transition times in periodically forced double-well potentials, the phenomenon of stochastic resonance, and the problem of approximating the metastable dynamics of a continuous-space system by a Markov chain on a finite set.
Based on joint works with Manon Baudel, Alexandra Blessing, Sébastien Dutercq, Bastien Fernandez, Barbara Gentz, Georgi Medvedev, Rita Nader, and Gideon Simpson.

