Searcharxiv⌕ Search

arXiv subjects

Takahiro Yoshida

Publications and source records attributed to Takahiro Yoshida.

32 records · Page 2Linked to original sources
arXiv2022

Subcritical regime of hybrid inflation with modular $A_4$ symmetry

We consider a supergravity model that has the modular $A_4$ symmetry and discuss the interplay between the neutrino mixing and inflation. The model contains right-handed neutrinos that have the Majorana masses and additional Yukawa couplings to the waterfall field. In the model an active neutrino is massless and we find that only the inverted hierarchy is allowed and the Majorana phase is predicted to be around $\pm (120\text{--}180)^\circ$ from the observed neutrino mixing data. In the early universe, one of right-handed sneutrinos plays the role of the inflaton field. Focusing on the subcritical regime of the hybrid inflation that is consistent with the cosmic microwave background data, we analyze the dynamics of the scalar sector and derive an upper bound $\mathcal{O}(10^{10})~{\rm GeV}$ on the scale of the Majorana mass.

hep-ph↗
arXiv2022

gwpcorMapper: an interactive mapping tool for exploring geographically weighted correlation and partial correlation in high-dimensional geospatial datasets

Exploratory spatial data analysis (ESDA) plays a key role in research that includes geographic data. In ESDA, analysts often want to be able to visualize observations and local relationships on a map. However, software dedicated to visualizing local spatial relations be-tween multiple variables in high dimensional datasets remains undeveloped. This paper introduces gwpcorMapper, a newly developed software application for mapping geographically weighted correlation and partial correlation in large multivariate datasets. gwpcorMap-per facilitates ESDA by giving researchers the ability to interact with map components that describe local correlative relationships. We built gwpcorMapper using the R Shiny framework. The software inherits its core algorithm from GWpcor, an R library for calculating the geographically weighted correlation and partial correlation statistics. We demonstrate the application of gwpcorMapper by using it to explore census data in order to find meaningful relationships that describe the work-life environment in the 23 special wards of Tokyo, Japan. We show that gwpcorMapper is useful in both variable selection and parameter tuning for geographically weighted statistics. gwpcorMapper highlights that there are strong statistically clear local variations in the relationship between the number of commuters and the total number of hours worked when considering the total population in each district across the 23 special wards of Tokyo. Our application demonstrates that the ESDA process with high-dimensional geospatial data using gwpcorMapper has applications across multiple fields.

stat.AP↗
arXiv2022

An Algorithm for Computing the Stratonovich's Value of Information

We propose an algorithm for computing Stratonovich's value of information (VoI) that can be regarded as an analogue of the distortion-rate function. We construct an alternating optimization algorithm for VoI under a general information leakage constraint and derive a convergence condition. Furthermore, we discuss algorithms for computing VoI under specific information leakage constraints, such as Shannon's mutual information (MI), $f$-leakage, Arimoto's MI, Sibson's MI, and Csiszar's MI.

cs.IT↗
arXiv2022

A Generalization of the Stratonovich's Value of Information and Application to Privacy-Utility Trade-off

The Stratonovich's value of information (VoI) is quantity that measure how much inferential gain is obtained from a perturbed sample under information leakage constraint. In this paper, we introduce a generalized VoI for a general loss function and general information leakage. Then we derive an upper bound of the generalized VoI. Moreover, for a classical loss function, we provide a achievable condition of the upper bound which is weaker than that of in previous studies. Since VoI can be viewed as a formulation of a privacy-utility trade-off (PUT) problem, we provide an interpretation of the achievable condition in the PUT context.

cs.IT↗
arXiv2021

Spatial prediction of apartment rent using regression-based and machine learning-based approaches with a large dataset

Employing a large dataset (at most, the order of n = 10^6), this study attempts enhance the literature on the comparison between regression and machine learning (ML)-based rent price prediction models by adding new empirical evidence and considering the spatial dependence of the observations. The regression-based approach incorporates the nearest neighbor Gaussian processes (NNGP) model, enabling the application of kriging to large datasets. In contrast, the ML-based approach utilizes typical models: extreme gradient boosting (XGBoost), random forest (RF), and deep neural network (DNN). The out-of-sample prediction accuracy of these models was compared using Japanese apartment rent data, with a varying order of sample sizes (i.e., n = 10^4, 10^5, 10^6). The results showed that, as the sample size increased, XGBoost and RF outperformed NNGP with higher out-of-sample prediction accuracy. XGBoost achieved the highest prediction accuracy for all sample sizes and error measures in both logarithmic and real scales and for all price bands (when n = 10^5 and 10^6). A comparison of several methods to account for the spatial dependence in RF showed that simply adding spatial coordinates to the explanatory variables may be sufficient.

stat.AP↗
arXiv2021

Modulus $τ$ linking leptonic CP violation to baryon asymmetry in $A_4$ modular invariant flavor model

We propose an $A_4$ modular invariant flavor model of leptons, in which both CP and modular symmetries are broken spontaneously by the vacuum expectation value of the modulus $τ$. The value of the modulus $τ$ is restricted by the observed lepton mixing angles and lepton masses for the normal hierarchy of neutrino masses at $3σ$ confidence level. The predictive Dirac CP phase $δ_{CP}$ is in the ranges $[0^\circ,50^\circ]$, $[170^\circ,175^\circ]$ and $[280^\circ,360^\circ]$ for ${\rm Re}\, [τ]<0$, and $[0^\circ,80^\circ]$, $[185^\circ,190^\circ]$ and $[310^\circ,360^\circ]$ for ${\rm Re}\,[τ]>0$ at $3σ$ confidence level. The sum of three neutrino masses is predicted in $[60, 84]$ meV, and the effective mass for the $0νββ$ decay is in [0.003, 3] meV. On the other hand, there is no allowed region of the modulus $τ$ for the inverted hierarchy of neutrino masses at $3σ$ confidence level. The modulus $τ$ links the Dirac CP phase to the cosmological baryon asymmetry (BAU) via the leptogenesis. Due to the strong wash-out effect, the predictive baryon asymmetry $Y_B$ can be at most the same order of the observed value. Then, the lightest right-handed neutrino mass is restricted in the range of $M_1 =[1.5,\,6.5] \times 10^{13}$ GeV. We find the correlation between the predictive $Y_B$ and the Dirac CP phase $δ_{CP}$. Only two predictive $δ_{CP}$ ranges, $[5^\circ,40^\circ]$ (${\rm Re}\,[τ]>0$) and $[320^\circ,355^\circ]$ (${\rm Re}\,[τ]<0$) are consistent with the BAU.

hep-ph↗
arXiv2021

Modular $A_4$ invariance and leptogenesis

We consider a model with three right-handed neutrinos in which Yukawa coupling constants and Majorana masses are obtained by requiring the modular $A_4$ symmetry. It has been shown that the model can explain mass hierarchies and mixing patterns of charged leptons and neutrinos with the seesaw mechanism. In this article we investigate the leptogenesis by decays of right-handed neutrinos in this model. It is shown that masses of right-handed neutrinos are about $10^{13}$ GeV in order to account for the observed baryon asymmetry of the universe. Furthermore, the positive sign of the baryon asymmetry is obtained only for the limited ranges of mixing angles and CP violation phases of active neutrinos, which can be tested by future neutrino experiments.

hep-ph↗
arXiv2021

Lepton flavor model with modular $A_4$ symmetry in large volume limit

We consider the modular symmetry associated with the compactification of extra dimensions as the flavor symmetry on lepton sector. Especially, we propose a model based on the modular $A_4$ symmetry with three right-handed neutrinos and a gauge singlet Higgs, which works well in the so-called large volume limit of the extra dimensions, i.e., $\mbox{Im}τ\to \infty$ for a modulus $τ$. The right-handed neutrinos are introduced to realize the seesaw mechanism for tiny neutrino masses observed in oscillation experiments. The vacuum expectation value of the singlet Higgs gives the Majorana masses for right-handed neutrinos and the $μ$-term for Higgs fields. The model can explain the observed masses and mixing angles of neutrinos successfully. We find that one of the mixing angles should be in the range $\sin^2 θ_{23} \ge 0.58$. Importantly, the CP violating phases of neutrinos are predicted in the two restricted regions. One is that the Dirac phase is $δ_{\rm CP} \simeq - 0.5 π$ and the Majorana phases are $α_{21} \simeq 0$ and $α_{31} \simeq π$. The other is $δ_{\rm CP} \simeq + 0.5 π$, $α_{21} \simeq 2 π$ and $α_{31} \simeq π$. The effective neutrino mass in the neutrinoless double beta decay is found to be $m_{\rm eff} = 0.037$--0.047 eV. These predictions will be tested in the future neutrino experiments.

hep-ph↗
arXiv2020

Scalable GWR: A linear-time algorithm for large-scale geographically weighted regression with polynomial kernels

Although a number of studies have developed fast geographically weighted regression (GWR) algorithms for large samples, none of them has achieved linear-time estimation, which is considered a requisite for big data analysis in machine learning, geostatistics, and related domains. Against this backdrop, this study proposes a scalable GWR (ScaGWR) for large datasets. The key improvement is the calibration of the model through a pre-compression of the matrices and vectors whose size depends on the sample size, prior to the leave-one-out cross-validation, which is the heaviest computational step in conventional GWR. This pre-compression allows us to run the proposed GWR extension so that its computation time increases linearly with the sample size. With this improvement, the ScaGWR can be calibrated with one million observations without parallelization. Moreover, the ScaGWR estimator can be regarded as an empirical Bayesian estimator that is more stable than the conventional GWR estimator. We compare the ScaGWR with the conventional GWR in terms of estimation accuracy and computational efficiency using a Monte Carlo simulation. Then, we apply these methods to a US income analysis. The code for ScaGWR is available in the R package scgwr. The code is embedded into C++ code and implemented in another R package, GWmodel.

stat.ME↗
arXiv2019

Resonant leptogenesis at TeV-scale and neutrinoless double beta decay

We investigate a resonant leptogenesis scenario by quasi-degenerate right-handed neutrinos which have TeV-scale masses. Especially, we consider the case when two right-handed neutrinos are responsible to leptogenesis and the seesaw mechanism for active neutrino masses, and assume that the CP violation occurs only in the mixing matrix of active neutrinos. In this case the sign of the baryon asymmetry depends on the Dirac and Majorana CP phases as well as the mixing angle of the right-handed neutrinos. It is shown how the yield of the baryon asymmetry correlates with these parameters. In addition, we find that the effective neutrino mass in the neutrinoless double beta decay receives an additional constraint in order to account for the observed baryon asymmetry depending on the masses and mixing angle of right-handed neutrinos.

hep-ph↗
arXiv2019

Distributed Stochastic Gradient Descent Using LDGM Codes

We consider a distributed learning problem in which the computation is carried out on a system consisting of a master node and multiple worker nodes. In such systems, the existence of slow-running machines called stragglers will cause a significant decrease in performance. Recently, coding theoretic framework, which is named Gradient Coding (GC), for mitigating stragglers in distributed learning has been established by Tandon et al. Most studies on GC are aiming at recovering the gradient information completely assuming that the Gradient Descent (GD) algorithm is used as a learning algorithm. On the other hand, if the Stochastic Gradient Descent (SGD) algorithm is used, it is not necessary to completely recover the gradient information, and its unbiased estimator is sufficient for the learning. In this paper, we propose a distributed SGD scheme using Low-Density Generator Matrix (LDGM) codes. In the proposed system, it may take longer time than existing GC methods to recover the gradient information completely, however, it enables the master node to obtain a high-quality unbiased estimator of the gradient at low computational cost and it leads to overall performance improvement.

cs.IT↗
arXiv2018

Which country epitomizes the world? A study from the perspective of demographic composition

Demographic indicators are an essential element in considering various problems in the social economy, such as predicting economic fluctuations and establishing policies. The literature widely discusses the growth of the world population or issues pertaining to its aging, but has given little to no attention to population structures and transition patterns. In this article, we take advantage of the characteristics of compositional data to examine the transition of the world population structure. Using the Aitchison distance, we examine the similarity of the world population structure from the 1990s to 2080 and that of countries and regions in 2015 and create maps to illustrate the results. Accordingly, we identify the following countries and regions as epitomes of the world population structure through different periods: India, Northern Africa and South Africa, in the 1990s, South America in 2015 to 2030, Oceania and Northern America in 2040, Uruguay and Puerto Rico in 2050 to 2060, and Italy and Japan in the distant future. We then cluster countries based on the similarity of their population structures in 2015 and correspond each cluster to a certain period. We found that Russia and Western Europe gather in a cluster that does not correspond to any period, indicating a recessive population structure.

stat.AP↗
arXiv2016

A Moran coefficient-based mixed effects approach to investigate spatially varying relationships

This study develops a spatially varying coefficient model by extending the random effects eigenvector spatial filtering model. The developed model has the following properties: its coefficients are interpretable in terms of the Moran coefficient; each of its coefficients can have a different degree of spatial smoothness; and it yields a variant of a Bayesian spatially varying coefficient model. Also, parameter estimation of the model can be executed with a relatively small computationally burden. Results of a Monte Carlo simulation reveal that our model outperforms a conventional eigenvector spatial filtering (ESF) model and geographically weighted regression (GWR) models in terms of the accuracy of the coefficient estimates and computational time. We empirically apply our model to the hedonic land price analysis of flood risk in Japan.

stat.ME↗