Institut de Physique Théorique Philippe Meyer


Home > Research > Statistical physics and biophysics

Takahiro Nemoto

by Marc-Thierry Jaekel - published on , updated on

Takahiro Nemoto joined the IPM in November 2016. He obtained his Ph.D. in March 2015 at Kyoto University, Japan, under the supervision of Shin.ichi Sasa. Before coming to the institute, he worked in ENS de Lyon with Freddy Bouchet from April 2015 to October 2015 and then in Université Paris-Diderot with Vivien Lecomte from November 2015 to October 2016.

Takahiro Nemoto’s research focuses on rare events in non-equilibrium systems. His expertise lies in developing sampling methods to accelerate the observation of rare events in numerical simulations (rare-event sampling methods) and in applying them to a variety of systems, such as tracer particles, non-equilibrium lattice gas models, kinetically constrained models for glassy dynamics, non-equilibrium active Brownian particles and turbulence. One of the mainstream ideas for rare event samplings is to use population dynamics in simulations. This idea has been developed and used by many groups, including Jorge Kurchan in LPENS who contributed to the establishment of a method to calculate large deviation functions in dynamical systems. The method involves simulations of system copies, where rare copies (or typical ones) are multiplied (or removed) during the simulations to efficiently sample target rare events. In order for the algorithm to work correctly, a certain number of copies is required, however, when the system shows singular behaviors, such as criticalities in dynamical phase transitions, this number rapidly increases, hampering the detailed analysis of these systems. To overcome this difficulty, Takahiro Nemoto with Freddy Bouchet, Robert Jack and Vivien Lecomte incorporated a control force within the algorithm by using a feedback scheme (similar to multi-canonical sampling methods) to decrease the required number of copies. The developed method has been successfully applied to the study of dynamical phase transitions in kinetically constrained models and also in active Brownian particles.

In the IPM, together with Alexandros Alexakis (LPENS), his main project is to develop rare event sampling methods to study turbulence. Their target is the prolonged time scale of laminar-turbulent transitions in pipe flows, which is related to the determination of the critical Reynolds number, initiated in 1883 by Osborne Reynolds. In his seminal paper, Reynolds asked at which speed (critical Reynolds number) does a laminar flow become turbulent? In a modern point of view, his question is summarized as such: by adding a small perturbation to a laminar flow in a pipe, we create a small turbulent patch called turbulent “puff”. This turbulent puff shows sudden decaying or splitting, whose time intervals are stochastic following an exponential (memory less) distribution function. The critical Reynolds number is then characterized as the point at which the decay and split time scales become equal. It was only in 2011 that this critical Reynolds number was measured for the first time in pipe flow experiments by Avila et al. However, the prolonged decay and split time scales do not allow the same investigation in numerical simulations (DNS) so far. To further understand these phenomena, Takahiro Nemoto with Alexandros Alexakis developed a rare event sampling method to facilitate the measurement of the puff decay time in numerical simulations. This simple method only relies on a feedback control of Reynolds number, which can be easily implemented in DNS. After publishing their fundamental idea, they applied to GENCI (Grand équipement national de calcul intensif) for access to HPC resources and obtained three million computational hours. Using them they are currently studying the time-series data obtained in DNS and applying their own method to it.

His published papers after his arrival at the IPM are [52-59].