セミナー

MNISQ: A Large-Scale Quantum Circuit Dataset for Machine Learning on/for Quantum Computers in the NISQ era

2023年8月29日(火) 14:00 - 15:30

Leonardo Placidi (大阪大学 大学院基礎工学研究科 博士課程)

We introduce the first large-scale dataset, MNISQ, for both the Quantum and the Classical Machine Learning community during the Noisy Intermediate-Scale Quantum era. MNISQ consists of 4,950,000 data points organized in 9 subdatasets. Building our dataset from the quantum encoding of classical information (e.g., MNIST dataset), we deliver a dataset in a dual form: in quantum form, as circuits, and in classical form, as quantum circuit descriptions (quantum programming language, QASM). In fact, also Machine Learning research related to quantum computers undertakes a dual challenge: enhancing machine learning by exploiting the power of quantum computers, while also leveraging state-of-the-art classical machine learning methodologies to help the advancement of quantum computing. Therefore, we perform circuit classification on our dataset, tackling the task with both quantum and classical models. In the quantum endeavor, we test our circuit dataset with Quantum Kernel methods, and we show excellent results with up to 97% accuracy. In the classical world, the underlying quantum mechanical structures within the quantum circuit data are not trivial. Nevertheless, we test our dataset on three classical models: Structured State Space sequence model (S4), Transformer, and LSTM. In particular, the S4 model applied on the tokenized QASM sequences reaches an impressive 77% accuracy. These findings illustrate that quantum circuit-related datasets are likely to be quantum advantageous, but also that state-of-the-art machine learning methodologies can competently classify and recognize quantum circuits. We finally entrust the quantum and classical machine learning community.

会場: 理化学研究所 和光キャンパス (メイン会場) / via Zoom

イベント公式言語: 英語

Landscape structure drives eco-evolution in host parasite systems

2023年8月24日(木) 16:00 - 17:00

Jhelam Deshpande (Ph.D. Student, Biodiversity: dynamics, interactions and conservation team, Institute of Evolutionary Science of Montpellier, France)

As all biological and many artificial systems, hosts and their parasites are most often spatially structured. Besides this highly relevant spatial context, parasites may change through time due to to evolutionary processes, including mutation and selection. These facts imply that we must study host-parasite systems taking into account space and evolution. Past work has mainly focused on simple spatial structures, but how parasites evolve in realistically complex landscapes remains unclear, hampering the translation of theoretical predictions to real ecological systems.Therefore, we here develop an eco-evolutionary metapopulation model of host-parasite interactions in which hosts and parasites disperse through realistically complex spatial graphs. Parasite virulence, a parasite life-history trait of central importance that here impacts host reproduction, is able to evolve. Our model therefore captures the eco-evolutionary feedback loop between host demography and parasite evolution in space. In order to gain a general understanding of parasite eco-evolution in space, we analyse our model for spatial networks that represent terrestrial (represented by random-geometric graphs; RGG) and riverine aquatic (represented by optimal channel networks; OCN) landscapes. We find that evolved virulence is generally a function of host dispersal, with a unimodal relationship in aquatic and a saturating relationship in terrestrial landscape, and this is driven by higher order network properies. Consistent with previous work, we show that our results are driven by kin selection, because dispersal and landscape structure impact both patterns of relatedness and availability of susceptible hosts. Our model yields readily testable predictions, including that terrestrial parasites should be more virulent than aquatic parasites are low dispersal rates and vice versa as dispersal increases. These differences in evolved virulence directly lead to differences in system stability, with more virulent parasites more often leading to host extinction. Thus, in this study we highlight the role of landscape structure in driving eco-evolutionary dynamics of parasites.

会場: via Zoom

イベント公式言語: 英語

Mating system of buckwheat

2023年8月17日(木) 16:00 - 17:00

ジェフリ・フォーセット (理化学研究所 数理創造プログラム (iTHEMS) 上級研究員)

Buckwheat (soba in Japanese) has a slightly unusual mating system called heterostylous self-incompatibility where two types of individuals coexist, one that produces flowers with a long style (female part of the flower) and short stamen (male part of the flower), and the other that produces flowers with a short style and long stamen. Mating is only successful when it occurs between the different types of individuals. It is a bit similar to sexual dimorphism where males and females coexists but in this case all individuals have both male and female organs. In this talk, I will introduce the basics of this mating system in buckwheat and some work we have been doing. In particular, I will talk about its genetic architecture and some parallels observed with other plants in which a similar mating system evolved independently. The talk will be aimed at non-experts so non-biologists are also welcome to attend.

会場: via Zoom

イベント公式言語: 英語

Dark matter heating vs vortex creep heating in old neutron stars

2023年8月7日(月) 13:30 - 15:00

藤原 素子 (Postdoctoral Researcher, Theoretical Particle Physics Group, Technical University of Munich, Germany)

Old isolated neutron stars have been gathering attention as targets to probe Dark Matter (DM) through temperature observations. DM will anomalously heat neutron stars through its gravitational capture and annihilation process, which predicts surface temperature as T_s ~ (1 − 3) × 10^3 K for t > 10^6 years. We may put constraints on DM-nucleon scattering cross section by finding even colder neutron stars. This story, however, assumed that there is no relevant heating source for old neutron stars. In this talk, we discuss the creep motion of vortex lines in the neutron superfluid of the inner crust as the heating mechanism. This creep mechanism is inherent in the structure of neutron stars. The heating luminosity is proportional to the time derivative of the angular velocity of the pulsar rotation, and the proportional constant J has an approximately universal value for each neutron star. If this vortex creep heating is quantitatively relevant against DM heating, this mechanism may cause a serious background to probe DM. The J parameter can be determined from the temperature observation of old neutron stars because the heating luminosity is balanced with the photon emission in the late time. We study the latest data of neutron star temperature observation and find that these data indeed give similar values of J, in favor of the assumption that these neutron stars are heated by the frictional motion of vortex lines. Besides, these values turn out to be consistent with the theoretical calculations of the vortex-nuclear interaction. Integarting all the results, we evaluate the vortex creep heating and conclude its significance against DM heating.

会場: セミナー室 (359号室) (メイン会場) / via Zoom

イベント公式言語: 英語

Quasi-local holography in 3d quantum gravity

2023年8月4日(金) 14:00 - 15:30

Etera Livine (Research Director CNRS, Ecole Normale Supérieure de Lyon, France)

Since the idea appeared in black hole physics, the concept of holography has become a guiding principle for quantum gravity. It is the notion that the dynamics of the geometry of a region of space-time can be entirely encoded in a theory living on its boundary. Although such holographic dualities have been well-developed in an asymptotical context, it remains a challenge to realize it exactly at finite distances. I will draw a possible route in 3d quantum gravity, by showing a duality between the Ponzano-Regge path integral for 3d quantum gravity as a topological field theory and the 2d (inhomogeneous) Ising model. This leads to an intriguing geometrical interpretation of the Ising critical couplings and opens the door to a possibly rich interplay between 3d quantum gravity and 2d condensed matter built out of holographic dualities.

会場: セミナー室 (359号室) (メイン会場) / via Zoom

イベント公式言語: 英語

Evidence against a strong first-order phase transition in neutron star cores: impact of new data

2023年8月1日(火) 13:30 - 15:00

Len Brandes (Ph.D. Student, Technical University of Munich, Germany)

Information on the phase structure of strongly interacting matter at high baryon densities can be gained from observations of neutron stars and their detailed analysis. Bayesian inference methods are used to set constraints on the speed of sound in the interior of neutron stars, based on recent multimessenger data in combination with low-density constraints based on chiral effective field theory and perturbative QCD constraints at asymptotically high densities. A detailed re-analysis is performed in order to clarify the influence of the latter constraints on the inference procedure. The impact of the recent new heavy (2.35 M_sol) black widow pulsar PSR J0952-0607 and of the unusually light supernova remnant HESS J1731-347 is inspected. One of the consequences of including PSR J0952-0607 in the database is a further stiffening of the equation-of-state, resulting in a 2.1 solar-mass neutron star in a reduced central density of less than five times the equilibrium density of normal nuclear matter. A systematic Bayes factor assessment quantifies the evidence (or non-evidence) for small sound speeds, necessary for a strong first-order phase transition, within the range of densities realized in the core of neutron stars. Given the presently existing database, it can be concluded that the occurrence of a strong first-order phase transition in the core of even a 2.1 solar-mass neutron star is unlikely, while a continuous crossover cannot be ruled out.

会場: via Zoom (メイン会場) / セミナー室 (132号室)

イベント公式言語: 英語

Clifford Group and Unitary Designs under Symmetry

2023年7月31日(月) 14:00 - 15:30

三橋 洋亮 (東京大学 大学院工学系研究科 物理工学専攻 博士課程)

We have generalized the well-known statement that the Clifford group is a unitary 3-design into symmetric cases by extending the notion of unitary design. Concretely, we have proven that a symmetric Clifford group is a symmetric unitary 3-design if and only if the symmetry constraint is described by some Pauli subgroup. We have also found a complete and unique construction method of symmetric Clifford groups with simple quantum gates for Pauli symmetries. For the overall understanding, we have also considered physically relevant U(1) and SU(2) symmetry constraints, which cannot be described by a Pauli subgroup, and have proven that the symmetric Clifford group is a symmetric unitary 1-design but not a 2-design under those symmetries. Our findings are numerically verified by computing the frame potentials, which measure the difference in randomness between the uniform ensemble on the symmetric group of interest and the symmetric unitary group. This work will open a new perspective into quantum information processing such as randomized benchmarking, and give a deep understanding to many-body systems such as monitored random circuits.

会場: 理化学研究所 和光キャンパス 研究本館3階 345-347 (メイン会場) / via Zoom

イベント公式言語: 英語

Exploring interior magnetic-field in neutron stars

2023年7月25日(火) 10:00 - 12:00

小嶌 康史 (広島大学 名誉教授)

Neutron stars are well known as a good laboratory to test high-density material. Magnetic field on the stars is relevant to some astrophysical phenomena, and understanding the effect is crucial to extract information of the stellar interior. In this talk, I start with discussing some evidence and implications for intense magnetic field, and focus on the magneto-elastic equilibrium in the solid crust. The study leads to upper limit of deformation, and beyond a threshold crustal fracture observed burst on strongly magnetized neutron stars. The model is still primitive, and further development is desirable to connect micro-physics with astrophysical observation.

会場: via Zoom (メイン会場) / 理化学研究所 和光キャンパス

イベント公式言語: 英語

Electronic instabilities emerging from higher-order van Hove singularities

2023年7月24日(月) 17:00 - 18:15

Xinloong Han (Postdoctoral Fellow, Kavli Institute for Theoretical Sciences, University of Chinese Academy of Sciences, China)

Time: 5pm ~ 6:15pm (JST); 10am ~ 11:15am (CET); 4pm ~ 5:15pm (Taiwan) Field: condensed matter physics Keywords: topological superconductor, Van Hove singularity, Hubbard model, Kagome lattices Abstract: Competing correlated electronic states are a central topic in condensed matter physics. A typical example is the close competition between spin density wave and d-wave superconductivity in the Hubbard model on the square lattice near half filling where the band structures have saddle points at which the Fermi surface topology changes from hole type to electron type. The saddle points are called van Hove singularity (VHS) points, and host diverging density of states with power-law behavior in the two dimensions. Recently, another type of VHS, namely the higher-order VHS was investigated in ABC-stacked trilayer graphene and twisted bilayer graphene. In this talk, I will first introduce the higher-order VHS, and make comparisons to the conventional VHS. Then I will discuss the enhanced nematicity driven by large flavor number with higher-order VHSs on the square and Kagome lattices. Finally, I will show that robust topological superconductivity can emerge on the square lattice due to interplay of spin-orbital coupling and higher-order VHSs.

会場: コモンルーム 246-248号室 とZoomのハイブリッド開催

イベント公式言語: 英語

Evolution of dormant egg production and their hatching rate in Aedes albopictus

2023年7月13日(木) 16:00 - 17:00

桑野 友輔 (総合研究大学院大学 博士課程)

Mosquitoes are important insect vectors of infectious diseases in humans, and knowledge of their population dynamics is pivotal in disease control. Some mosquito species have dormancy in their life history to survive harsh environments. However, the population dynamics of mosquitoes have not yet been well understood due to the lack of field and experimental data on dormancy. For that reason, I modeled the population dynamics of mosquitoes that face environmental fluctuations and examine the evolution of egg dormancy strategy to survive harsh periods. I found that the ESS dormancy fraction monotonically increases with the period of environmental fluctuation. Next, I analyzed evolutionary traits of the dependence of the dormancy rate and the hatching rate from dormant egg on soil moisture content and conducted evolutionary simulations using actual weather measurement in Tokyo. The results of the hatching rate from dormant egg showed that two mosquito phenotypes having distinctly different responses to soil moisture were selected.

会場: via Zoom

イベント公式言語: 英語

Conserved charges in the quantum simulation of integrable spin chains

2023年7月12日(水) 13:30 - 15:00

Juan William Pedersen (東京大学 大学院総合文化研究科 博士課程)

In this talk, we present the result of the quantum simulation of the spin-1/2 Heisenberg XXX spin chain. We implement the integrable Trotterization algorithm, which allows us to control the Trotter error with conserved charges remaining conserved, on a real quantum computer and classical simulators. We study the effects of quantum noise on the time evolution of several conserved charges and specifically observe the decay of the expectation values. Our work improves our understanding of quantum noises and can potentially be applied to benchmark quantum devices and algorithms.

会場: セミナー室 (359号室) (メイン会場) / via Zoom

イベント公式言語: 英語

Searching for dark matter subhalos in the Fermi-LAT catalog with Bayesian neural networks

2023年7月10日(月) 16:30 - 18:00

Slivia Manconi (Marie Skłodowska-Curie Fellow, Laboratoire d'Annecy-Le-Vieux de Physique Theorique (LAPTh), CNRS, France)

Machine learning techniques are powerful tools to tackle diverse tasks in current astroparticle physics research. For example, Bayesian neural networks provide robust classifiers with reliable uncertainty estimates, and are particularly well suited for classification problems that are based on comparatively small and imbalanced data sets, such as the gamma-ray sources detected by Fermi-Large Area Telescope (LAT). About one third of the gamma-ray sources collected in the most recent catalogs remain currently unidentified. Intriguingly, some of these could be exotic objects such as dark subhalos, which are overdensities in dark matter halos predicted to form by cosmological N-body simulations. If they exist in the Milky Way, they could be detected as gamma-ray point sources due to the annihilation or decay of dark matter particles into Standard Model final states. In this talk I will discuss our recent work* in which, after training on realistic simulations, we use Bayesian neural networks to identify candidate dark matter subhalos among unidentified gamma-ray sources in Fermi-LAT catalogs. Our novel framework allows us to derive conservative bounds on the dark matter annihilation cross section, by excluding unidentified sources classified as astrophysical-like.

会場: via Zoom

イベント公式言語: 英語

A dynamical model for IRAS 00500+6713: the remant of a type Iax supernova SN1181 hosting a double degenerate merger product WD J005311

2023年7月7日(金) 14:00 - 15:15

黄 天鋭 (東京大学 ビッグバン宇宙国際センター (RESCEU) 博士課程)

Iras 00500+6713 is a bright nebula in the infrared, and X-ray observations show it consists of diffuse region and strong illuminated central region. In addition, optical spectral observations have recently revealed that fast wind with about 15,000 km/s is blowing from the massive white dwarf at the center. The properties of this nebula and white dwarf are very similar to those theoretically predicted by the binary white dwarf merger. In addition, its position on the celestial sphere and the extent make it a prime candidate for the remnant of SN 1181, a historical supernova. In this study, we propose that such a multilayered structure is formed by the collision between the remnant of SN 1181 and the stellar wind blowing from the central white dwarf, and succeeded in constructing a model that is consistent with the multi-wavelength observations. The results show that the progenitor of SN 1181 is a binary white dwarf with 1.3-1.9 solar mass and that their merger triggered an explosion that ejected mass with 0.2-0.6 solar mass to form the present object. The extent of the X-ray source concentrated in the center reveals that these winds began blowing within the last 30 years, and we will discuss this property as well.

会場: セミナー室 (359号室) / via Zoom

イベント公式言語: 英語

Introduction to braid groups

2023年7月5日(水) 14:00 - 16:30

根上 春 (千葉大学 大学院融合理工学府数学情報科学専攻 博士課程)

Part 1 (14:00-15:00): Introduction to braid groups Braid groups are groups that are defined by figures formed by the entanglement of n strings. Besides this geometric realization, it is a very interesting field where algebra and analysis intersect. In the first half of this seminar, aimed mainly at those unfamiliar with braid groups, we will introduce three aspects of braid groups and review the history of the research. In particular, in the area of its relation to analysis, the relationship between KZ equations and braid groups will be introduced. Part 2 (15:30-16:30): Representations of braid groups and the relationship between monodromy representations of KZ equations In the second half of the talk, after a brief introduction to representation theory, we will introduce the Katz-Long-Moody construction, a method of constructing infinite series of representations of the semi-direct product of braid group and free group. We will also show that its special case is isomorphic to multiplicative middle convolution, a method for constructing monodromy representations of KZ equations. Lastly, we will also discuss the connection between representations of braid groups and knot invariants. The talk includes joint work with Kazuki Hiroe.

会場: セミナー室 (359号室) / via Zoom

イベント公式言語: 英語

Maximum Force Conjecture and Black Hole Thermodynamics

2023年6月30日(金) 14:00 - 15:30

Yen Chin Ong (Professor, Center for Gravitation and Cosmology, College of Physical Science and Technology, Yangzhou University, China)

I review the current controversial status of the so-called "maximum force conjecture" in general relativity, whose validity has recently been debated. Then I will discuss how maximum force conjecture can nevertheless be relevant for black hole thermodynamics.

会場: 研究本館 3階 359号室とZoomのハイブリッド開催

イベント公式言語: 英語

Journal Club: Inference of hydrodynamic equations for active matter

2023年6月29日(木) 16:00 - 17:00

足立 景亮 (理化学研究所 生命機能科学研究センター (BDR) 生体非平衡物理学理研白眉研究チーム 基礎科学特別研究員)

イベント公式言語: 英語

The classical equations of motion of quantised gauge theories

2023年6月23日(金) 13:30 - 15:00

トム・メリア (東京大学 カブリ数物連携宇宙研究機構 (Kavli IPMU) 准教授)

The Einstein and Maxwell equations are the jewels in the crown of classical physics. But classical physics is only an approximation to nature, arising as a limit of the underlying quantum mechanical description. And in the case of both general relativity and electromagnetism, owing to their gauge theory nature, the full set of classical equations of motion are not guaranteed to follow from the quantum theory. The time-time and time-space components of the Einstein equations in GR and Gauss’ law in EM are enforced ‘by hand' in the quantisation procedure—a choice so as to make the classical-like states behave as per our classical belief. But what if our universe was actually described by another classical-like state? For GR, the resulting modification of the Einstein equations can be packaged as the inclusion of an auxiliary energy-momentum tensor describing a ’shadow’ matter that adds no additional degrees of freedom to the theory. The homogeneous and isotropic background piece of this auxiliary matter contributes to expansion of the universe identical to cold dark matter, and the inhomogeneous components source curvature perturbations that grow linearly at linear order.

会場: 研究本館 3階 359号室とZoomのハイブリッド開催

イベント公式言語: 英語

The role of fluid dynamics in microbial ecology

2023年6月22日(木) 14:00 - 15:00

Douglas R. Brumley (Senior Lecturer, Applied Mathematics, University of Melbourne, Australia)

Bacterial motility, symbioses, and marine nutrient cycling unfold at the scale of individual microbes, and are inherently dynamic. In this talk, I will discuss the role that fluid flows play in shaping the ecology of microbes, both in the open ocean as well as around coral surfaces. In each case, I will demonstrate how iteratively combining video-microscopy, image processing and mathematical modelling can resolve features of microbial lifestyles that are difficult or impossible to see otherwise, and show how single-cell measurements can be connected to bulk processes at the population-level.

会場: 研究本館 3階 359号室とZoomのハイブリッド開催

イベント公式言語: 英語

Matrix estimation via singular value shrinkage

2023年6月21日(水) 15:30 - 16:30

松田 孟留 (理化学研究所 脳神経科学研究センター (CBS) 統計数理連携ユニット ユニットリーダー)

In this talk, I will introduce recent studies on shrinkage estimation of matrices. First, we develop a superharmonic prior for matrices that shrinks singular values, which can be viewed as a natural generalization of Stein’s prior. This prior is motivated from the Efron–Morris estimator, which is an extension of the James–Stein estimator to matrices. The generalized Bayes estimator with respect to this prior is minimax and dominates MLE under the Frobenius loss. In particular, since it shrinks to the space of low-rank matrices, it attains large risk reduction when the unknown matrix is close to low-rank (e.g. reduced-rank regression). Next, we construct a theory of shrinkage estimation under the “matrix quadratic loss”, which is a matrix-valued loss function suitable for matrix estimation. A notion of “matrix superharmonicity” for matrix-variate functions is introduced and the generalized Bayes estimator with respect to a matrix superharmonic prior is shown to be minimax under the matrix quadratic loss. The matrix-variate improper t-priors are matrix superharmonic and this class includes the above generalization of Stein’s prior. Applications include matrix completion and nonparametric estimation.

会場: 研究本館 3階 359号室とZoomのハイブリッド開催

イベント公式言語: 英語