A core mission of the group is developing theories in stochastic and quantum thermodynamics for small nonequilibrium systems. We study how energy dissipation, information, and fluctuations govern the behavior of mesoscopic machines — from biochemical oscillators and autonomous information engines to open quantum systems. Recent advances extend our framework from classical stochastic thermodynamics into the quantum regime, addressing relaxation control and thermodynamic trade-off relations.

Representative results are given as follows:

Accelerating quantum relaxation via temporary reset: a Mpemba-inspired approach

Slow relaxation in open quantum systems limits the performance of quantum heat engines and computation devices. Inspired by the Mpemba effect, we proposed a protocol that temporarily couples the system to a reset channel, enabling significant acceleration of relaxation from any initial state. Unlike previous approaches targeting a single mode, our method can simultaneously suppress multiple relaxation modes, providing a versatile tool for controlling quantum thermodynamic timescales.

Phys. Rev. Lett. 135, 150403 (2025).

Universal trade-off between irreversibility and intrinsic timescale in thermal relaxation

We established a universal trade-off relation between irreversibility (entropy production) and the intrinsic relaxation timescale during thermal relaxation processes. This relation sets fundamental bounds on how fast a system can relax given a fixed dissipation budget, with direct applications to thermodynamic inference — estimating entropy production from experimentally accessible observables without detailed knowledge of the underlying dynamics.

Phys. Rev. E (2025).

Designing autonomous Maxwell's demon via stochastic resetting

Autonomous Maxwell's demons are information engines that produce work by exploiting a memory tape. We showed that stochastic resetting can dramatically enhance their performance: it drives the demon to its functional periodic steady state at the fastest pace and expands its effective work regions. We also discovered a dual-function region in the demon's phase diagram where it simultaneously produces work and erases information, resolving this apparent paradox through a modified Clausius inequality.

Phys. Rev. Research 5, 043066 (2023).

Design principles for biochemical oscillations with limited energy resources

Biochemical systems may frequently suffer from limited energy resources. We derive the energy-accuracy and sensitivity-accuracy trade-off relations for a general biochemical model, finding that systems have a much lower energy cost by using noise-induced oscillations to keep almost equal efficiency compared with normal oscillations. An optimal system size is predicted where both the highest sensitivity and accuracy can be reached simultaneously.

Phys. Rev. Research, 2, 043331 (2020).

Mode-coupling theory study of glassy dynamics in active particles system

We presented a mode coupling theory study for the relaxation and glassy dynamics of a system of strongly interacting self-propelled active particles. We find the critical density shifts to larger values with increasing magnitude of propulsion force or effective temperature, in good accordance with simulation. Our theory recovers the results for passive systems and can be extended to more complex systems such as active-passive mixtures.

Soft Matter, 13, 4464 (2017).

Fluctuation effect in a genetic toggle switch

We discovered that a conventional biological genetic toggle switch can exhibit an extra stable state near the deterministically unstable fixed point. Stochastic nullcline analysis reveals that the smallness and discreteness of critical molecule numbers are responsible, demonstrating a fundamental role of intrinsic noise in shaping the phenotypic landscape of gene regulatory circuits.

Phys. Rev. Lett. 109, 248107 (2012).