Events

Date: 23 January 2025   Time: 13:00 - 14:00     

Location: MB-503

Machine learning methods have recently been developed for computationally intensive investigations of rare events in nonequilibrium systems. However, present methods have generally been designed for Markov processes, presenting a major limitation due to the important role of memory in many realistic models. Examples of memory dependence are routinely encountered in complex systems, where models include hidden variables. Here we introduce a reinforcement learning method for non-Markov systems that extends the actor-critic framework given by Rose et al. [New.J.Phys.23, 013013(2021)] for obtaining scaled cumulant generating functions characterizing the fluctuations. The actor-critic is implemented using neural-networks; a particular innovation in our method is the use of an additional neural policy for processing memory-variables . We demonstrate results for current fluctuations in various memory-dependent models, with a special focus on semi-Markov systems, where hidden variables control the inter-event waiting time distributions.