The spacecraft Hayabusa has returned to the earth in 2010. The event was closed up and widely broadcast because the spacecraft succeeded in returning back tiny particles from the asteroid Itokawa. Preliminary examination team of Hayabusa returned samples investigated the small particles around a few tens μm, and found a lot of facts about the past evolution and future evolution of the small asteroid. Challenges on the analysis of the Hayabusa returned samples were to maximize the scientific gain from 40 tiny samples with a coordinated analytical flow with the sequences of various analytical methods. Mounting of the small precious samples on the holders is always problematic because the sample was damaged or even lost during such procedures. Moreover, accuracy of the analysis largely depends on such sample preparation procedure. In the analysis of the Hayabusa returned samples, the team applied the sequence which was used for the analysis of a previous sample return project, though they could not investigate the samples with atmosphere shielded environment, for all samples. The Hayabusa2 spacecraft, a next generation spacecraft for sample return missions, successfully touched down on the asteroid Ryugu, which will contain water and organic materials. Thus, it should be much difficult to analyze the samples than that of the Hayabusa returned samples because those water and organic materials are easily damaged and contaminated by terrestrial materials. In this paper, we report the processes of trial and error for the analysis of the Hayabusa returned samples, and preparation for the analysis of the Hayabusa2 returned samples based on the learning by the trial and error.

Basic concept and few advanced topics of deep learning are reviewed. We explain supervised machine learning, neural nets, optimization with SGD, theoretical reason to make network deep, adversarial nets, an open question on generalization, and deep learning libraries.

Fundamental problems on the hydrophobic effect and the hydrophobic interaction are reviewed. It is remarked that the solvation thermodynamics characteristic of hydrophobic hydration is actually not so uncommon when the constant-volume solvation process is considered while it is indeed atypical in the constant-pressure process, and what makes water special as a solvent of hydrophobic species is the smallness of its thermal pressure coefficient. What we know so far about the hydrophobic interaction are also described. Recent theoretical studies demonstrate that the osmotic second virial coefficient of methane in water is positive at low temperatures, decreases monotonically with temperature, and is large and negative at high temperatures, thus urging caution in assuming the hydrophobic interaction to be always attractive. A mean-field approximation for the solvation free energy in liquid mixtures is reviewed, which gives simple physical pictures on microscopic mechanisms of the temperature, pressure, and salt concentration dependences of the solvation free energy and the hydrophobic interaction.

We study the unconventional quantum critical behavior observed in the Yb-based quasicrystal and its approximant crystal. We show that (i) both the quasicrystal and approximant crystal have the same critical exponent and (ii) the divergent phenomenon in the magnetic susceptibility is also observed under pressure with the same critical exponent in the quasicrystal, whereas the approximant exhibits quantum critical behavior only at the critical pressure. We discuss the possible origin of the unconventional quantum criticality.

A quantum walk, that is, a synthetic quantum system whose dynamics is described by a time-evolution operator, provides potential applications for quantum computing and information as well as quantum simulators. It is further interesting that the quantum walk possesses novel topological phases akin to those of Floquet topological insulators. In this article, we present how to study topological phases on a nonunitary quantum walk with parity and time-reversal symmetry. Accordingly, we show that the quantum walk provides a potential application for controlling of probability amplitudes of photons.