arXiv subjects

Hiroki Sakamoto

Publications and source records attributed to Hiroki Sakamoto.

14 recordsLinked to original sources

arXiv2026

A Deep State-Space Model Compression Method using Upper Bound on Output Error

We study deep state-space models (Deep SSMs) that contain linear quadratic-output (LQO) systems as internal blocks and present a compression method with a provable output error guarantee. We first derive an upper bound on the output error between two Deep SSMs and show that the bound can be expressed in terms of the $h^2$-error norms between the layerwise LQO systems. In particular, we show that reducing the $h^2$ approximation errors of the LQO systems placed in shallow layers is effective in reducing the derived upper bound on the output error. Next, we formulate an optimization problem for the derived upper bound and develop a gradient-based MOR method. In the numerical experiments, using the IMDb task from the LRA benchmark, we demonstrate the effectiveness of the proposed upper-bound-based compression method. In particular, we show that the number of trainable parameters can be reduced by approximately 60\% without retraining while maintaining the performance of the original model.

eess.SY↗

arXiv2026

Data-Driven Regularized Time-Limited h2 Model Reduction from Noisy Impulse Responses

This paper develops a data-driven time-limited h2 model reduction method for discrete-time linear time-invariant systems. Specifically, we formulate and solve a regularized time-limited h2 model reduction problem using only noisy impulse response data. Furthermore, we show that the objective function and its gradient can be represented using only noisy impulse response data. Numerical experiments using SLICOT benchmarks demonstrate that the proposed regularized method achieves lower relative time-limited h2 errors than the tested alternatives and is effective in situations where the unregularized method may deteriorate under noise.

eess.SY↗

arXiv2025

Data-driven h2 model reduction for linear discrete-time systems

We present a data-driven framework for $h^{2}$-optimal model reduction for linear discrete-time systems. Our main contribution is to create optimal reduced-order models in the $h^{2}$-norm sense directly from the measurement data alone, without using the information about the original system. In particular, we focus on the fact that the gradients of the $h^{2}$ model reduction problem are expressed using the discrete-time Lyapunov equation and the discrete-time Sylvester equation, and derive the data-driven gradients. The proposed algorithm uses the output of an existing MOR as the initial point, and convergence to a stationary point is guaranteed under certain assumptions. In numerical experiments, we demonstrate that, for a modeling task in neuroscience, our method constructs a reduced-order model that outperforms DMDc in terms of the $h^{2}$-norm.

math.OC↗

arXiv2025

Compression Method for Deep Diagonal State Space Model Based on $H^2$ Optimal Reduction

Deep learning models incorporating linear SSMs have gained attention for capturing long-range dependencies in sequential data. However, their large parameter sizes pose challenges for deployment on resource-constrained devices. In this study, we propose an efficient parameter reduction method for these models by applying $H^{2}$ model order reduction techniques from control theory to their linear SSM components. In experiments, the LRA benchmark results show that the model compression based on our proposed method outperforms an existing method using the Balanced Truncation, while successfully reducing the number of parameters in the SSMs to $1/32$ without sacrificing the performance of the original models.

cs.LG↗

arXiv2025

Neutron star in Logarithmic model of Cartan $F(R)$ gravity

Cartan $F(R)$ gravity introduces the equivalent scalar-tensor theory by extending the gravity sector. From the solution of the modified Cartan equation leads to the interaction with the scalar field and the fermion. We derived the effective potential of the scalar field using the auxiliary field method, which is commonly used in studies of spontaneous chiral symmetry breaking. Using this effective potential we investigated the mass-radius relation of neutron star to solve the Tolman-Oppenheimer-Volkoff equation. By the numerical computation, we found that the contribution of the scalar field increases the mass of a neutron star. This could give theoretical plausibility to the current observations.

gr-qc↗

arXiv2024

ARIM-mdx Data System: Towards a Nationwide Data Platform for Materials Science

In modern materials science, effective and high-volume data management across leading-edge experimental facilities and world-class supercomputers is indispensable for cutting-edge research. However, existing integrated systems that handle data from these resources have primarily focused just on smaller-scale cross-institutional or single-domain operations. As a result, they often lack the scalability, efficiency, agility, and interdisciplinarity, needed for handling substantial volumes of data from various researchers. In this paper, we introduce ARIM-mdx data system, aiming at a nationwide data platform for materials science in Japan. Currently in its trial phase, the platform has been involving 11 universities and institutes all over Japan, and it is utilized by over 800 researchers from around 140 organizations in academia and industry, being intended to gradually expand its reach. The ARIM-mdx data system, as a pioneering nationwide data platform, has the potential to contribute to the creation of new research communities and accelerate innovations.

cs.DC↗

arXiv2023

Robustness of predicted CMB fluctuations in Cartan $F(R)$ gravity

The cosmology of the $F(R)$ gravity rebuilding by the Cartan formalism is investigated. This is called Cartan $F(R)$ gravity. The well-known $F(R)$ gravity has been introduced to extend the standard cosmology, e.g. to explain the cosmological accelerated expansion as the inflation. Cartan $F(R)$ gravity is based on the Riemann-Cartan geometry. The curvature $R$ can separate to two parts, one is derived from the Levi-Civita connection and the other from the torsion. Assuming the matter-independent spin connection, we have successfully rewritten the action of Cartan $F(R)$ gravity into the Einstein-Hilbert action and a scalar field with canonical kinetic and potential terms without any conformal transformations. This feature simplifies building and analysis of new model of inflation. In this paper, we study two models, the power-law model and logarithmic model, and evaluate fluctuations in the cosmological microwave background (CMB) radiation. We found the robustness of CMB fluctuation by the analytical computation and confirm this feature by the numerical calculation.

gr-qc↗

arXiv2023

Stable Linear System Identification with Prior Knowledge by Riemannian Sequential Quadratic Optimization

We consider an identification method for a linear continuous time-invariant autonomous system from noisy state observations. In particular, we focus on the identification to satisfy the asymptotic stability of the system with some prior knowledge. To this end, we propose to model this identification problem as a Riemannian nonlinear optimization (RNLO) problem, where the stability is ensured through a certain Riemannian manifold and the prior knowledge is expressed as nonlinear constraints defined on this manifold. To solve this RNLO, we apply the Riemannian sequential quadratic optimization (RSQO) that was proposed by Obara, Okuno, and Takeda (2022) most recently. RSQO performs quite well with theoretical guarantee to find a point satisfying the Karush-Kuhn-Tucker conditions of RNLO. In this paper, we demonstrate that the identification problem can be indeed solved by RSQO more effectively than competing algorithms.

math.OC↗

arXiv2019

Exploring the inflation of $F(R)$ gravity

The early time expansion of the space-time, namely inflation, is introduced to solve some cosmological problems. $F(R)$ gravity is a simple extension of the general relativity to induce the inflationary expansion. The precise observation of the Cosmic Microwave Background gives us the information to inspect the model of the inflation. By the Weyl transformation, $F(R)$ gravity can be transformed to the Einstein-Hilbert term with a scalar field that plays the role of the inflaton. In this paper we obtain general formulae to derive the inflationary parameters including the models with a non-canonical kinetic term. The inflationary parameters are described as functions of the inflation potential and its derivatives. We apply the procedure to $F(R)=R+γR^n$ and $R(1+γR^n)^{1/n}$ models and show the allowed parameters space.

gr-qc↗

arXiv2018

On some parabolic systems arising from a nuclear reactor model with nonlinear boundary conditions

In this paper, we are concerned with a reaction diffusion system arising from a nuclear reactor model in bounded domains with nonlinear boundary conditions. We show the existence of a stationary solution and its ordered uniqueness. It is also shown that every positive stationary solution possesses threshold property to determine blow-up or globally existence for solutions of nonstationary problem.

math.AP↗

arXiv2018

Phenomenological Aspects of Possible Vacua of a Neutrino Flavor Model

We discuss a supersymmetric model with discrete flavor symmetry $A_4\times Z_3$. The additional scalar fields which contribute masses of leptons in the Yukawa terms are introduced in this model. We analyze their scalar potential and find that they have various vacuum structures. We show the relations among 24 different vacua and classify them into two types. We derive expressions of the lepton mixing angles, Dirac CP violating phase and Majorana phases for the two types. The model parameters which are allowed by the experimental data of the lepton mixing angles are different for each type. We also study the constraints on the model parameters which are related to Majorana phases. The different allowed regions of the model parameters for the two types are shown numerically for a given region of two combinations of the CP violating phases.

hep-ph↗

arXiv2017

An alternative attractor in gauged NJL inflation

We have investigated the attractor structure for the CMB fluctuations in composite inflation scenario within the gauged Nambu-Jona-Lasinio (NJL) model. Such composite inflation represents an attractor which can not be found in a fundamental scalar model. As is known, the number of inflationary models contains the attractor classified by the $α$-attractor model. It is found that the attractor inflation in the gauged NJL model corresponds to the $α= 2$ case.

hep-th↗

arXiv2017

Inflation from the Finite Scale Gauged Nambu-Jona-Lasinio Model

The possibility to construct an inflationary universe scenario for the finite-scale gauged Nambu-Jona-Lasinio model is investigated. This model can be described by the Higgs-Yukawa type interaction model with the corresponding compositeness scale. Therefore, the one-loop Higgs-Yukawa effective potential is used with the compositeness condition for the study of inflationary dynamics. We evaluate the fluctuations in the cosmic microwave background for the model with a finite compositeness scale in the slow-roll approximation. We find the remarkable dependence on the gauge group and the number of fermion flavors. It is also proved that the model has similar behavior with the $ϕ^{4n}$ chaotic inflation and the Starobinsky model at the flat and steep limits, respectively. It is demonstrated that realistic inflation consistent with Planck data is possible for a range of theory parameters.

hep-ph↗

arXiv2015

Gauged Nambu-Jona-Lasinio inflation

We investigate the gauged Nambu-Jona-Lasinio model in curved spacetime at the large $N_c$ limit and in slow-roll approximation. The model can be described by the renormalization group corrected gauge-Higgs-Yukawa theory with the corresponding compositeness conditions. Evaluating the renormalization group (RG) improved effective action, we show that such model can produce CMB fluctuations and find inflationary parameters: spectral index, tensor-to-scalar-ratio and running of the spectral index. We demonstrate that the model can naturally satisfy the Planck 2015 data and maybe considered as an alternative candidate for Higgs inflation.

hep-th↗