Convergence acceleration by non-reversibility and degenerate noise (B05)

• Subject Area: Mathematics
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 539309657

This project aims to develop a systematic quantitative theory of convergence to stationarity for non-reversible Markov processes with a focus on high-dimensional state spaces and degenerate noise. In particular, we will consider the acceleration of Markov chain Monte Carlo methods, convergence acceleration in the absence of convexity, and the role of non-reversibility for stochastic dynamics at and near criticality.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 1720: Analysis of criticality: from complex phenomena to models and estimates
• Applicant Institution: Rheinische Friedrich-Wilhelms-Universität Bonn
• Project Head: Professor Dr. Andreas Eberle, since 7/2025