Attaining entropy production and dissipation maps from Brownian movies via neural networks
This provides a practical tool for researchers studying stochastic systems at mesoscopic scales, such as biological assemblies, to uncover nonequilibrium properties, though it is an incremental application of existing neural network techniques to a specific domain.
The researchers tackled the problem of quantifying entropy production (EP) from experimental time-series image data without tracking relevant variables, by developing a convolutional neural network (CNN) method that accurately measures EP and produces spatiotemporal dissipation maps, even with noisy or low-resolution data.
Quantifying entropy production (EP) is essential to understand stochastic systems at mesoscopic scales, such as living organisms or biological assemblies. However, without tracking the relevant variables, it is challenging to figure out where and to what extent EP occurs from recorded time-series image data from experiments. Here, applying a convolutional neural network (CNN), a powerful tool for image processing, we develop an estimation method for EP through an unsupervised learning algorithm that calculates only from movies. Together with an attention map of the CNN's last layer, our method can not only quantify stochastic EP but also produce the spatiotemporal pattern of the EP (dissipation map). We show that our method accurately measures the EP and creates a dissipation map in two nonequilibrium systems, the bead-spring model and a network of elastic filaments. We further confirm high performance even with noisy, low spatial resolution data, and partially observed situations. Our method will provide a practical way to obtain dissipation maps and ultimately contribute to uncovering the nonequilibrium nature of complex systems.