[注意事項]
発表の10日前までに office[at]ap.t.utokyo.ac.jp 宛てに「氏名」「指導教員」「発表題目（英語）」「要旨（英語）」「発表言語（英語または日本語）」を送付して下さい。
 発表日
 2021年10月22日（金）16：50〜
Aグループ
 座長
 小林 弘和
 指導
教員名
 渡辺 悠樹 准教授

 座長
 小山 奏汰
 指導
教員名
 長谷川 達生 教授

発表者名 
石﨑 雄士 
指導教員名 
福谷 克之 教授 
発表題目（英語） 
Observation of the Edelstein Effect using a SpinPolarized Atomic Hydrogen Beam. 
要旨（英語） 
Nowadays, spin current is actively studied because of its characteristic behavior in topological insulators or topological superconductors and it is required to improve the efficiency of spin current injection[1]. While the spin Hall effect is commonly used as a means of spin current detection, the Edelstein Effect (EE) has recently attracted attention for the spin current measurement[2]. When charge current flows in the spinpolarized system, spin accumulation occurs, which is called the EE. The Inverse Edelstein Effect (IEE) is also observed[3], which means spin current generates charge current. We have been developing a spinpolarized atomic hydrogen beam, which emits proton with spinpolarized electron, as a tool to observe surface states[4]. We expect that the IEE can be induced, where the spinpolarized atomic hydrogen beam causes a charge current. For this purpose, we are considering to study the Au(111) surface. Au(111) has a 2D electron system localized on the surface. Because of relativistic effect, Rashba effect emerges to the surface electronic states of Au(111) and the spin splitting occurs. Recently, it is discovered that this states also can be recognized as topological states, which called Topological Derived Surface State(TDSS)[5]. Due to spin interaction of the spinpolarized hydrogen, it is expected a charge current is induced at the Rashbasplit surface. I will talk about the prediction of the effect on the TDSS and the plan of the experiment.
[1]L. Liu and C.F. Pai, et al.: Science 336 (2012) 555.
[2]V. M. Edelstein: Solid State Commun. 73 (1990) 233.
[3] J. C. Rojas Sánchez, et al.: Nat. Commun. 4, (2013) 2944.
[4]S. Ogura et al.: J. Vac. Soc. Jpn. 54 (2011) 192.
[5]B. Yan, et al.: Nat. Commun. 6 (2015) 10167. 
発表言語 
日本語 

発表者名 
宇都 隆宏 
指導教員名 
志村 努 教授 
発表題目（英語） 
Orbital angular momentum dichroism in plasmonic nanostructure 
要旨（英語） 
Interaction between materials with chirality, such as enantiomers, and light with circular polarisation, also denoted as spin angular momentum (SAM), has been researched since the 19th century. Recently in addition to SAM, orbital angular momentum (OAM) has been gaining attention as a new degree of freedom.
In this presentation, I will explain the background of these studies, especially the interaction between the LaguerreGaussian beam, which has OAM, and plasmonic nanostructure with chirality. I will also introduce how to measure OAM dichroism experimentally. 
発表言語 
日本語 

発表者名 
大島 久典 
指導教員名 
渡辺 悠樹 准教授 
発表題目（英語） 
Measurementindeuced entanglement phase transition in symmetric random unitary circuits 
要旨（英語） 
Quantum circuits subject to repeated projective measurements can exhibit an entanglement phase transition as a function of the rate of measurements, stemming from the competition between the scrambling effects of unitary dynamics and the disentangling effects of measurements. However, it is not understood exactly what kind of model will exhibit what phase and is characterized by what order parameters in general, and also the universality of the transition.
I will give a brief summary on the entanglement phase transition and talk about an idea that the symmetry of the model might be the key to understanding it in general like that of conventional theory in thermodynamics.

発表言語 
日本語 

Bグループ
 座長
 財津 英明
 指導
教員名
 岩佐 義宏 教授

 座長
 阪本 天志
 指導
教員名
 吉岡 孝高 准教授

発表者名 
魚住 亮介 
指導教員名 
吉岡 孝高 准教授 
発表題目（英語） 
Development of a highresolution probing laser suited for cold positronium Doppler spectroscopy 
要旨（英語） 
Cooling of positronium (Ps), an electronpositron bound state, down to a temperature below 10 K leads to important research such as precise measurements of energy intervals or realizing BoseEinstein condensation. Laser cooling and other cooling methods are on intense research to achieve this goal. On this occasion, evaluating the temperature of cold Ps is an important task to be undertaken. Obtaining the Doppler broadening by laser spectroscopy is one of the preferred methods since it allows highresolution measurements regarding the laser technology of today.
To conduct efficient and highresolution laser spectroscopy, the probing laser should have unique frequency structures along with sufficient laser intensity. As for the frequency structure, the linewidth of the laser should be optimized between the balance of high resolution and quantity of Ps at resonance, while the longitudinal mode interval should be comparable to the natural width of Ps (50 MHz) to efficiently excite Ps inside the laser linewidth. Adequate laser intensity would also be a demand to excite enough Ps to detect it as a signal. A numerical simulation based on optical Bloch equations, with full consideration of the unique frequency structure, would verify whether the laser has sufficient intensity.
Since these requirements turned out to be unachievable with commercially available lasers, a frequencytunable 243nm pulsed laser has been developed. This laser fulfills the demanding frequency structures, and results of numerical simulation have shown supportive proof that the current laser intensity is sufficient. 
発表言語 
日本語 

発表者名 
梅村 洸介 
指導教員名 
齊藤 英治 教授 
発表題目（英語） 
Time and spaceresolved spectroscopy of nonlinear spin waves 
要旨（英語） 
Spin waves, elementary excitations of magnetization, have strong nonlinearities caused by dipole and exchange interactions. This nonlinearity is a source of phase correlation between different modes, which give rise to a standingwave pattern of spin waves in real space. In the previous studies, the magnon nonlinearity has been investigated by microwave spinwave spectroscopy and Brillouin light scattering (BLS), while the spatiotemporal observation of the nonlinear magnon process has yet to be demonstrated.
In my presentation, I will give an overview of the spinwave spectroscopy measurement system we have constructed and show a movie of nonlinear spinwave excitation obtained by the measurement system. 
発表言語 
日本語 

発表者名 
大竹 雄太郎 
指導教員名 
求 幸年 教授 
発表題目（英語） 
Study on learning of deep neural networks using spin glass theory 
要旨（英語） 
Deep learning with deep neural networks (DNN) has many applications in modern society [1]. With proper training, the DNN will exhibit a high accuracy for unknown data. However, there are many unknowns about why networks can have the high generalization performance.
By using the replica method for DNNs and solving the saddle point equations for a kRSB ansatz, it is suggested that when α becomes large (α is defined by M = αN, where M is the number of data and N is the width of the network), the spin glass transition occurs from the layer close to the input/output boundary of the network [2].
On the other hand, in actual learning, a loss function is defined and the parameters in the network are updated to reduce the value of the loss. It is also useful to investigate the change of spin glass order parameter during the actual learning process. This will lead to new insights into deep learning and effective guidelines on networks for new problems.
In this presentation, I will introduce the results I have obtained so far on what happens inside DNN during learning.
[1] Y. LeCun, Y. Bengio and G. Hinton, Deep learning, Nature 521, 436 (2015)
[2] H. Yoshino, SciPost Phys. Core 2, 005 (2020) 
発表言語 
日本語 
