Prediction of Thermal Conductance of Complex Networks with Deep Learning

Abstract

Predicting thermal conductance of complex networks poses a formidable challenge in the field of materials science and engineering. This challenge arises due to the intricate interplay between the parameters of network structure and thermal conductance, encompassing connectivity, network topology, network geometry, node inhomogeneity, and others. Our understanding of how these parameters specifically influence heat transfer performance remains limited. Deep learning offers a promising approach for addressing such complex problems. In this Letter, we find the wellestablished convolutional neural network models AlexNet can predict the thermal conductance of complex network effciently. Our approach further optimizes the calculation effciency by reducing the image recognition in consideration that the thermal transfer is inherently encoded within the Laplacian matrix. Intriguingly, our findings reveal that adopting a simpler convolutional neural network architecture can achieve a comparable prediction accuracy while requiring less computational time. This result facilitates a more effcient solution for predicting the thermal conductance of complex networks and serves as a reference for machine learning algorithm in related domains.

Machine Learning Aided Design and Optimization of Thermal Metamaterials

“…In essence, the thermal transport behavior of a complex network is decisively determined by its Laplacian matrix when its topology is presented . Considering that deep learning algorithms typically operate on image data sets, which are converted into RGB-colored matrices (tensors in PyTorch), it is anticipated that replacing image data with the Laplacian matrix as inputs for predicting thermal transport properties may enhance the neural network’s understanding of heat flow mechanisms within complex network structures, as recently reported by Zhu et al Extending this philosophy, one may deduce that mechanical response information is encoded within the dynamical matrix . Therefore, it is conceivable to predict the mechanical performance of a complex network by training a deep learning model with the dynamical matrix.…”

“…In essence, the thermal transport behavior of a complex network is decisively determined by its Laplacian matrix when its topology is presented. 384 Considering that deep learning algorithms typically operate on image data sets, which are 385 Extending this philosophy, one may deduce that mechanical response information is encoded within the dynamical matrix. 386 Therefore, it is conceivable to predict the mechanical performance of a complex network by training a deep learning model with the dynamical matrix.…”

Structural Characterization of Horizontal Visibility Network Based on Time Series of Emergency Disease Visits

“…For example, medical scientists study the correlation between various diseases (or the relationship between the effects of drugs) through complex networks 9,10 . Physicists explore how physical phenomena manifest on complex networks [11][12][13][14] . Sociologists analyze interpersonal relationships and social structures through the construction of social networks 15,16 .…”