Energy-landscape design principle of smart molecules for temporal pattern recognition.

In contrast to molecules that are equilibrated in stationary environments, molecules can be driven out of equilibrium by a temporally changing environment. When a molecule has a complex energy landscape and a wide spectrum of relaxation rates, the information of the temporal pattern can be recorded into its nonequilibrium kinetics and the transient probability distribution of meta-stable configurations. In this work, we examine the energy-landscape design principle of smart molecules that can recognize and respond to different temporal patterns of a changing environment. By constructing a minimal graph model to capture the change of energy landscape at various environmental conditions, we optimize the design of a molecule capable of distinguishing environmental temporal patterns. This allows us to study the tradeoff relation between pattern recognition performance, molecule’s complexity, and the operational cost in terms of entropy production.