Speaker
Chun-Biu Li
(SU)
Description
A thermodynamic framework of entropy and information
transfers at the ensemble level has been recently formulated
(Horowitz et al. PRX 4:031015, 2014) for bipartite Markovian
systems, where the dynamical variables of the composite
subsystems can change only one at a time. Motivated by the
fact that bipartite Markovian systems are ubiquitous in
biomolecular systems, e.g., in the chemo-mechanical coupling
of motor proteins, etc., we generalize consistently the
thermodynamics of information flow to the stochastic
trajectory level in terms of notions of stochastic
thermodynamics. This allows us to unveil details of the
information processes, namely, entropy production, heat
dissipation, information flow, etc., among the composite
subsystems that were masked at the ensemble level. Moreover,
I will re-exam the interpretation of information flow, a
concept commonly used in the literature, by making an
explicit comparison to the well-known concept of transfer
entropy.
