Our iTHEMS member, Dr. Akinori Tanaka (AIP/iTHEMS) together with Dr. Akio Tomiya (RIKEN BNL Research Center) have received "14th Particle Physics Medal: Young Scientist Award in Theoretical Particle Physics” from Japan Particle and Nuclear Theory Forum. They were awarded in recognition of their recent paper: Akinori Tanaka, Akio Tomiya “Detection of Phase Transition via Convolutional Neural Networks,” J. Phys. Soc. Jpn. 86, 063001 (2017).
The citation reads: This paper answers the important question of whether machine learning can capture the existence of phase transition. In this study, in the classical Ising model, spin configurations at various temperatures are generated, and pairs of (configuration, temperature) are used as data. A convolutional neural network is prepared, input data is set, output data is temperature, and machine learning is performed to predict the temperature from the configuration. In the obtained neural network, when the weight of the network connecting the final layer and the layer immediately before it is plotted as a heat map, it can be seen that a large change occurs near the phase transition temperature. This is evidence that machine learning has captured a phase transition. This paper, which shows that machine learning can not only reproduce and use known concepts but also extract the physical concept of phase transition point, is very original and links the subsequent physics and machine learning It made one of the foundations of research. From these facts, we consider this paper to be an award-winning paper.
(The above citation was translated from the original in Japanese by use of Google Translation, in order to demonstrate the recent development of Machine Learning)

Congratulations, Akinori and Akio!