arXiv subjects

Masahiro Kobayashi

Publications and source records attributed to Masahiro Kobayashi.

At least 19 recordsLinked to original sources

arXiv2026

Contextual Bandit-Based Decomposition of Network Slice Requirements under Cumulative Resource Budget Constraints

End-to-end (E2E) network slices (NSs) are provisioned across multiple domains of the 5G network. In hierarchical NS management, a tenant submits a network slice request (NSR), which specifies E2E service level agreement (SLA) requirements. Rather than managing these domains directly, an E2E controller decomposes each NSR into domain-level SLA requirements and delegates resource allocation to domain-specific controllers, which return feasibility and resource-consumption feedback. A poor decomposition policy can therefore cause rejection of the current request by producing infeasible requirements or reduce future admission opportunities by concentrating resource consumption in bottleneck domains. We call this decomposition-policy optimization problem the network slice request decomposition problem (NSR-DP). For practical operation, online approaches to NSR-DP have been proposed. Such approaches must jointly meet two requirements: (R1) control long-term resource budgets and (R2) adapt each decomposition to the performance targets and guarantee levels specified in the arriving NSR's SLA. To meet these requirements, we introduce contextual constrained kernel bandits (CCKB) as an online solution for NSR-DP. To address (R1), CCKB raises penalties for using resources that become tight, thereby discouraging decompositions that consume bottleneck resources. To address (R2), it uses Gaussian processes (GPs) to predict, for the current NSR, the reward and resource usage of candidate decompositions, allowing it to select a decomposition suited to the performance targets and guarantee levels. We establish high-probability guarantees for the resulting formulation and show through extensive 5G simulations across topology, bottleneck, and traffic-mixture settings that CCKB outperforms the baselines in the large majority of conditions.

cs.NI↗

arXiv2026

Heavy-traffic limit of stationary distributions of a state-dependent queue

Inspired by the work of Atar and Miyazawa [1] (2026) as well as applications to energy-saving problems, we are interested in the heavy-traffic limit of the stationary queue length distribution, which is not addressed in [1]. In this paper, we consider this heavy-traffic limit for the single server queue which has the most general possible state-dependence. Namely, arrival and service speeds may take any values depending on the queue length. Here, the terminology, heavy-traffic limit, stands for a diffusion-scaled limit in heavy-traffic for processes, distributions and modeling primitives. This general model is referred to as a state-dependent queue. There are two motivations for this generalization. One is interest in the state-dependent queue itself because it allows finer control of service speed in application. Another is making it clear how the heavy-traffic limit is obtained under what conditions for the state-dependent queue. Thus, we start to study basic properties of this state-dependent queue, including its stability. We then take the sequence of the stationary distributions of its diffusion scaled queue-length processes. We have three main results for this sequence. We first show that it is tight if the heavy-traffic limit of their drifts exists and is negative as the queue length goes to infinity, where a drift is the arrival speed minus the service speed. We next assume the condition that the limit of every vaguely convergent subsequence has a density, which is referred to as a density condition, and show that the heavy-traffic limit of the stationary distributions is obtained in a closed form if and only if that negative drift condition holds. We then show that the density condition is always satisfied for the multi-level queue, so the problem is nicely solved for the multi-level queue.

math.PR↗

arXiv2026

Diffusion limit for the stationary distribution of a history-dependent two-level M/M/1 queue

Recently, Atar and Miyazawa [2] introduced a multi-level GI/G/1 queue with a finite number of levels, where both the arrival and service rates depend on the level corresponding to the current queue length. For this model, they proved that the diffusion limit of its queue length process in heavy traffic is the level-dependent reflected Brownian motion of [6]. In a subsequent study, Kobayashi et al. [4] derived the corresponding diffusion limit of the stationary distribution. These studies are motivated by the control of service capacity depending on the queue length. We are interested in the more general case where this control may also depend on the history of the queue length. As the first step toward such a generalization, we specialize the multi-level GI/G/1 queue to a two-level M/M/1 queue. We then extend the dynamics of this model so that its arrival and service rates depend not only on the current queue length but also on the recent history of queue lengths. Under the stability condition for this model, we first compute its stationary distribution in closed form, then derive its diffusion limit in heavy traffic. Finally, using this diffusion limit, we derive approximation formulas for the stationary distribution and then numerically assess their accuracy.

math.PR↗

arXiv2025

Heavy traffic limit of stationary distribution of the multi-level single server queue

Atar and Miyazawa recently introduced a single server queue with queue length dependent arrival and service processes, and name it a multi-level queue. They prove that the heavy traffic limit of its queue length process weakly converges to a reflected diffusion with discontinuously state-dependent drift and deviations. We derive the heavy traffic limit of the stationary queue length distribution of this multi-level queue in a closed form, which agrees with the stationary distribution of the reflected diffusion obtained by Miyazawa (2024, Journal of the Indian Society for Probability and Statistics). Thus, those results show the limit interchange of process and stationary distribution in heavy traffic. The multi-level queue is a simpler version of the 2-level GI/G/1 queue of Miyazawa (2025, Advances in Applied Probability, to appear) and its extension for multi-levels. For this 2-level queue in heavy traffic, the process limit is unknown, and the distributional limit is obtained for limited cases under extra conditions. Nevertheless, it is shown that the method developed in Miyazawa (2025) perfectly works for the present multi-level queue.

math.PR↗

arXiv2025

Relationship between Hölder Divergence and Functional Density Power Divergence: Intersection and Generalization

In this study, we discuss the relationship between two families of density-power-based divergences with functional degrees of freedom -- the Hölder divergence and the functional density power divergence (FDPD) -- based on their intersection and generalization. These divergence families include the density power divergence and the $γ$-divergence as special cases. First, we prove that the intersection of the Hölder divergence and the FDPD is limited to a general divergence family introduced by Jones et al. (Biometrika, 2001). Subsequently, motivated by the fact that Hölder's inequality is used in the proofs of nonnegativity for both the Hölder divergence and the FDPD, we define a generalized divergence family, referred to as the $ξ$-Hölder divergence. The nonnegativity of the $ξ$-Hölder divergence is established through a combination of the inequalities used to prove the nonnegativity of the Hölder divergence and the FDPD. Furthermore, we derive an inequality between the composite scoring rules corresponding to different FDPDs based on the $ξ$-Hölder divergence. Finally, we prove that imposing the mathematical structure of the Hölder score on a composite scoring rule results in the $ξ$-Hölder divergence.

cs.IT↗

arXiv2025

A Unified Representation of Density-Power-Based Divergences Reducible to M-Estimation

Density-power-based divergences are known to provide robust inference procedures against outliers, and their extensions have been widely studied. A characteristic of successful divergences is that the estimation problem can be reduced to M-estimation. In this paper, we define a norm-based Bregman density power divergence (NB-DPD) -- density-power-based divergence with functional flexibility within the framework of Bregman divergences that can be reduced to M-estimation. We show that, by specifying the function $ϕ_γ$, NB-DPD reduces to well-known divergences, such as the density power divergence and the $γ$-divergence. Furthermore, by examining the combinations of functions $ϕ_γ$ corresponding to existing divergences, we show that a new divergence connecting these existing divergences can be derived. Finally, we show that the redescending property, one of the key indicators of robustness, holds only for the $γ$-divergence.

cs.IT↗

arXiv2024

Experimental Validation of Collision-Radiation Dataset for Molecular Hydrogen in Plasmas

Quantitative spectroscopy of molecular hydrogen has generated substantial demand, leading to the accumulation of diverse elementary-process data encompassing radiative transitions, electron-impact transitions, predissociations, and quenching. However, their rates currently available are still sparse and there are inconsistencies among those proposed by different authors. In this study, we demonstrate an experimental validation of such molecular dataset by composing a collisional-radiative model (CRM) for molecular hydrogen and comparing experimentally-obtained vibronic populations across multiple levels. From the population kinetics of molecular hydrogen, the importance of each elementary process in various parameter space is studied. In low-density plasmas (electron density $n_\mathrm{e} \lesssim 10^{17}\;\mathrm{m^{-3}}$) the excitation rates from the ground states and radiative decay rates, both of which have been reported previously, determines the excited state population. The inconsistency in the excitation rates affects the population distribution the most significantly in this parameter space. On the other hand, in higher density plasmas ($n_\mathrm{e} \gtrsim 10^{18}\;\mathrm{m^{-3}}$), the excitation rates \textit{from} excited states become important, which have never been reported in the literature, and may need to be approximated in some way. In order to validate these molecular datasets and approximated rates, we carried out experimental observations for two different hydrogen plasmas; a low-density radio-frequency (RF) heated plasma ($n_\mathrm{e}\approx 10^{16}\;\mathrm{m^{-3}}$) and the Large Helical Device (LHD) divertor plasma ($n_\mathrm{e}\gtrsim 10^{18}\;\mathrm{m^{-3}}$)... [continued]

physics.plasm-ph↗

arXiv2024

Unbiased Estimating Equation on Inverse Divergence and Its Conditions

This paper focuses on the Bregman divergence defined by the reciprocal function, called the inverse divergence. For the loss function defined by the monotonically increasing function $f$ and inverse divergence, the conditions for the statistical model and function $f$ under which the estimating equation is unbiased are clarified. Specifically, we characterize two types of statistical models, an inverse Gaussian type and a mixture of generalized inverse Gaussian type distributions, to show that the conditions for the function $f$ are different for each model. We also define Bregman divergence as a linear sum over the dimensions of the inverse divergence and extend the results to the multi-dimensional case.

cs.IT↗

arXiv2023

Plasma-Parameter Dependence of Ro-Vibrational Temperatures for $\mathrm{H}_2$ in LHD Divertor

We analyzed a thousand visible spectra of Fulcher-$α$ band measured for divertor plasmas in Large Helical Device. With a coronal model and Baysian inference, the population distribution of hydrogen molecule in the electronical ground state were estimated. The non-thermal population distribution was recovered with a two-temperature model, which has two sets of rotational and vibrational temperatures, as well as their mixture coefficient. The lower rotational temperature significantly changes according to the plasma parameters. Its nearly linear dependence on the electron density was found, which is consistent with previous works. The lower vibrational temperature also shows a small density dependence, as reported by a previous work. On the other hand, the higher rotational and vibrational temperatures as well as the mixture coefficient only show slight changes over the broad range of plasma parameters. These population parameters show a significant correlation; with higher electron density, all the temperatures and the fraction of the higher-temperature component increase simultaneously. This suggests that the electron-impact plays an important role to determine the population distribution.

physics.plasm-ph↗

arXiv2023

Overfitting in quantum machine learning and entangling dropout

The ultimate goal in machine learning is to construct a model function that has a generalization capability for unseen dataset, based on given training dataset. If the model function has too much expressibility power, then it may overfit to the training data and as a result lose the generalization capability. To avoid such overfitting issue, several techniques have been developed in the classical machine learning regime, and the dropout is one such effective method. This paper proposes a straightforward analogue of this technique in the quantum machine learning regime, the entangling dropout, meaning that some entangling gates in a given parametrized quantum circuit are randomly removed during the training process to reduce the expressibility of the circuit. Some simple case studies are given to show that this technique actually suppresses the overfitting.

quant-ph↗

arXiv2021

Generalized Dirichlet-process-means for $f$-separable distortion measures

DP-means clustering was obtained as an extension of $K$-means clustering. While it is implemented with a simple and efficient algorithm, it can estimate the number of clusters simultaneously. However, DP-means is specifically designed for the average distortion measure. Therefore, it is vulnerable to outliers in data, and can cause large maximum distortion in clusters. In this work, we extend the objective function of the DP-means to $f$-separable distortion measures and propose a unified learning algorithm to overcome the above problems by selecting the function $f$. Further, the influence function of the estimated cluster center is analyzed to evaluate the robustness against outliers. We demonstrate the performance of the generalized method by numerical experiments using real datasets.

cs.LG↗

arXiv2020

Unbiased Estimation Equation under $f$-Separable Bregman Distortion Measures

We discuss unbiased estimation equations in a class of objective function using a monotonically increasing function $f$ and Bregman divergence. The choice of the function $f$ gives desirable properties such as robustness against outliers. In order to obtain unbiased estimation equations, analytically intractable integrals are generally required as bias correction terms. In this study, we clarify the combination of Bregman divergence, statistical model, and function $f$ in which the bias correction term vanishes. Focusing on Mahalanobis and Itakura-Saito distances, we provide a generalization of fundamental existing results and characterize a class of distributions of positive reals with a scale parameter, which includes the gamma distribution as a special case. We discuss the possibility of latent bias minimization when the proportion of outliers is large, which is induced by the extinction of the bias correction term.

cs.LG↗

arXiv2020

Multi-Decoder RNN Autoencoder Based on Variational Bayes Method

Clustering algorithms have wide applications and play an important role in data analysis fields including time series data analysis. However, in time series analysis, most of the algorithms used signal shape features or the initial value of hidden variable of a neural network. Little has been discussed on the methods based on the generative model of the time series. In this paper, we propose a new clustering algorithm focusing on the generative process of the signal with a recurrent neural network and the variational Bayes method. Our experiments show that the proposed algorithm not only has a robustness against for phase shift, amplitude and signal length variations but also provide a flexible clustering based on the property of the variational Bayes method.

stat.ML↗

arXiv2019

Extendable NFV-Integrated Control Method Using Reinforcement Learning

Network functions virtualization (NFV) enables telecommunications service providers to realize various network services by flexibly combining multiple virtual network functions (VNFs). To provide such services, an NFV control method should optimally allocate such VNFs into physical networks and servers by taking account of the combination(s) of objective functions and constraints for each metric defined for each VNF type, e.g., VNF placements and routes between the VNFs. The NFV control method should also be extendable for adding new metrics or changing the combination of metrics. One approach for NFV control to optimize allocations is to construct an algorithm that simultaneously solves the combined optimization problem. However, this approach is not extendable because the problem needs to be reformulated every time a new metric is added or a combination of metrics is changed. Another approach involves using an extendable network-control architecture that coordinates multiple control algorithms specified for individual metrics. However, to the best of our knowledge, no method has been developed that can optimize allocations through this kind of coordination. In this paper, we propose an extendable NFV-integrated control method by coordinating multiple control algorithms. We also propose an efficient coordination algorithm based on reinforcement learning. Finally, we evaluate the effectiveness of the proposed method through simulations.

cs.DC↗

arXiv2018

Exact asymptotic formulae of the stationary distribution of a discrete-time 2d-QBD process: an example and additional proofs

A discrete-time two-dimensional quasi-birth-and-death process (2d-QBD process), $\{\boldsymbol{Y}_n\}=\{(X_{1,n},X_{2,n},J_n)\}$, is a two-dimensional skip-free random walk $\{(X_{1,n},X_{2,n})\}$ on $\mathbb{Z}_+^2$ with a supplemental process $\{J_n\}$ on a finite set $S_0$. The supplemental process $\{J_n\}$ is called a phase process. The 2d-QBD process $\{\boldsymbol{Y}_n\}$ is a Markov chain in which the transition probabilities of the two-dimensional process $\{(X_{1,n},X_{2,n})\}$ vary according to the state of the phase process $\{J_n\}$. This modulation is assumed to be space homogeneous except for the boundaries of $\mathbb{Z}_+^2$. Under certain conditions, the directional exact asymptotic formulae of the stationary distribution of the 2d-QBD process have been obtained in "T. Ozawa and M. Kobayashi, Exact asymptotic formulae of the stationary distribution of a discrete-time two-dimensional QBD process, Queueing Systems (2018) DOI:10.1007/s11134-018-9586-x." In this paper, we give an example of 2d-QBD process and proofs of some lemmas and propositions appeared in that paper.

math.PR↗

arXiv2017

Exact asymptotic formulae of the stationary distribution of a discrete-time two-dimensional QBD process

We consider a discrete-time two-dimensional process $\{(L_{1,n},L_{2,n})\}$ on $\mathbb{Z}_+^2$ with a supplemental process $\{J_n\}$ on a finite set, where individual processes $\{L_{1,n}\}$ and $\{L_{2,n}\}$ are both skip free. We assume that the joint process $\{Y_n\}=\{(L_{1,n},L_{2,n},J_n)\}$ is Markovian and that the transition probabilities of the two-dimensional process $\{(L_{1,n},L_{2,n})\}$ are modulated depending on the state of the background process $\{J_n\}$. This modulation is space homogeneous except for the boundaries of $\mathbb{Z}_+^2$. We call this process a discrete-time two-dimensional quasi-birth-and-death (2D-QBD) process and, under several conditions, obtain the exact asymptotic formulae of the stationary distribution in the coordinate directions.

math.PR↗

arXiv2016

Simple error bounds for the QBD approximation of a special class of two dimensional reflecting random walks

This paper considers the QBD approximation of a special class of two-dimensional reflecting random walks (2D-RRWs). A typical example of the 2D-RRWs is a two-node Jackson network with cooperative servers. The main contribution of this paper is to provide simple upper bounds for the relative absolute difference between the time-averaged functionals of the original 2D-RRW and its QBD approximation.

math.PR↗

arXiv2014

Tail asymptotics of the stationary distribution of a two dimensional reflecting random walk with unbounded upward jumps

We consider a two dimensional reflecting random walk on the nonnegative integer quadrant. This random walk is assumed to be skip free in the direction to the boundary of the quadrant, but may have unbounded jumps in the opposite direction, which are referred to as upward jumps. We are interested in the tail asymptotic behavior of its stationary distribution, provided it exists. Assuming the upward jump size distributions have light tails, we completely find the rough tail asymptotics of the marginal stationary distributions in all directions. This generalizes the corresponding results for the skip free reflecting random walk in Miyazawa (2009) [Mathematics of Operations Research 34, 547-575]. We exemplify these results for a two node network with exogenous batch arrivals.

math.PR↗