Memory-Driven Theory Links Consecutive Rare Events
A new analytical framework shows that long-term memory in stochastic systems creates correlations between successive rare events and shifts the waiting-time distribution away from exponential, breaking the Arrhenius-Kramers paradigm. Biswas and Guérin's generalized Langevin approach reveals how power-law memory slows relaxation, producing event clustering and enabling parameter-free predictions of the full first-passage-time distribution from equilibrium measurements. The work has implications for climate risk, earthquakes, finance, and single-molecule experiments, offering a predictive theory of extreme events beyond memoryless models.
