TOHOKU UNIVERSITY Researchers

Details of the Researcher

Home

日本語 English

Kazuyuki Tanaka

Section

Graduate School of Information Sciences

Job title

Professor

Degree

工学博士（東北大学）

researchmap

https://researchmap.jp/Kazuyuki_Tanaka

J-GLOBAL ID

200901076140730739

Research History 6

2007/04 - Present

Tohoku University　Graduate School of Information Sciences　Professor
1999/04 - 2007/03

Tohoku University　Graduate School of Information Sciences　Associate Professor
1994/07 - 1999/03

Muroran Institute of Technology　Department of Computer Science and Systems Engineering　Associate Professor
1997/03 - 1998/02

University of Glasgow　Department of Statistics　Honorary Research Fellow
1993/06 - 1994/06

Tohoku University　Graduate School of Information Sciences　Assistant Professor
Tohoku University　Faculty of Engineering　Assistant Professor

Show all Show first 5

Education 3

Tohoku University　Graduate School, Division of Engineering　電子工学専攻

- 1989/03/24
Tohoku University　Graduate School, Division of Engineering　電子工学専攻

- 1986/03/25
Tohoku University　Faculty of Engineering　電子工学科

- 1984/03/27

Committee Memberships 26

Grant-in Aid for Scientific Research on Priority Areas: Statistical Mechanical Approach to Probabilistic Information Processing　Head Investigator

2002/04 - 2005/03
文部科学省特定領域研究「確率的情報処理への統計力学的アプローチ」　領域代表

2002/04 - 2005/03
電子情報通信学会　和文論文誌 (D-II)「情報論的学習理論」特集号編集委員

2004/03 - 2005/02
電子情報通信学会　和文論文誌 (D-II)「情報論的学習理論」特集号編集委員

2004/03 - 2005/02
電子情報通信学会東北支部　庶務幹事

2001/04 - 2003/03
電子情報通信学会東北支部　庶務幹事

2001/04 - 2003/03
電子情報通信学会　和文論文誌 (D-II)「情報論的学習理論」特集号編集委員

2001/07 - 2002/03
電子情報通信学会　和文論文誌 (D-II)「情報論的学習理論」特集号編集委員

2001/07 - 2002/03
2005 年情報論的学習理論ワークショップ (IBIS2005)　実行委員

2005/11 -
IEEE Computational Intelligence Society Japan Chapter　運営委員

2005/11 -
2005 年情報論的学習理論ワークショップ (IBIS2005)　実行委員

2005/11 -
IEEE Computational Intelligence Society Japan Chapter　運営委員

2005/11 -
Satellite Meeting of STATPHYS 22 Statistical Physics of Disordered Systems and Its Applications　組織委員

2005/07 -
Satellite Meeting of STATPHYS 22 Statistical Physics of Disordered Systems and Its Applications　組織委員

2005/07 -
電子情報通信学会情報論的学習理論時限研究専門委員会　専門委員

2004/11 -
電子情報通信学会情報論的学習理論時限研究専門委員会　専門委員

2004/11 -
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「確率的アルゴリズムによる情報処理」　世話人代表

2003/12 -
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「確率的アルゴリズムによる情報処理」　世話人代表

2003/12 -
2003 年情報論的学習理論ワークショップ (IBIS2003)　実行委員

2003/11 -
2003 年情報論的学習理論ワークショップ (IBIS2003)　実行委員

2003/11 -
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「確率的情報処理としての移動体通信技術」　世話人代表

2002/12 -
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「確率的情報処理としての移動体通信技術」　世話人代表

2002/12 -
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「ベイズ統計と統計力学を用いた確率的情報処理技術」　世話人代表

2002/03 -
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「ベイズ統計と統計力学を用いた確率的情報処理技術」　世話人代表

2002/03 -
日本物理学会 2000年春の分科会領域 11 シンポジューム「情報処理技術としての統計力学」　世話人代表

2000/03 -
日本物理学会 2000年春の分科会領域 11 シンポジューム「情報処理技術としての統計力学」　世話人代表

2000/03 -

Show all ︎Show first 5

Professional Memberships 1

The Physical Society of Japan

Research Interests 8

statistical science
mathematical science
Bayesian network
massive statistical model
probabilistic inference
image processing
statistical method
probabilistic information processing

Research Areas 4

Natural sciences / Mathematical physics and basic theory / Statistial Physics
Informatics / Intelligent informatics /
Informatics / Statistical science /
Informatics / Information theory / Mathematical Science

Awards 1

インターネットアーキテクチャ研究賞

2011/06/16　電子情報通信学会通信ソサイエティインターネットアーキテクチャー研究会

Papers 157

Assessment of image generation by quantum annealer Peer-reviewed

Takehito Sato, Masayuki Ohzeki, Kazuyuki Tanaka

Scientific Reports　11　(1)　13523-1-13523-8　2021/12
Publisher: Springer Science and Business Media LLC
DOI： 10.1038/s41598-021-92295-9 　

eISSN： 2045-2322

More details Close

<title>Abstract</title>Quantum annealing was originally proposed as an approach for solving combinatorial optimization problems using quantum effects. D-Wave Systems has released a production model of quantum annealing hardware. However, the inherent noise and various environmental factors in the hardware hamper the determination of optimal solutions. In addition, the freezing effect in regions with weak quantum fluctuations generates outputs approximately following a Gibbs–Boltzmann distribution at an extremely low temperature. Thus, a quantum annealer may also serve as a fast sampler for the Ising spin-glass problem, and several studies have investigated Boltzmann machine learning using a quantum annealer. Previous developments have focused on comparing the performance in the standard distance of the resulting distributions between conventional methods in classical computers and sampling by a quantum annealer. In this study, we focused on the performance of a quantum annealer as a generative model from a different aspect. To evaluate its performance, we prepared a discriminator given by a neural network trained on an a priori dataset. The evaluation results show a higher performance of quantum annealer compared with the classical approach for Boltzmann machine learning in training of the generative model. However the generation of the data suffers from the remanent quantum fluctuation in the quantum annealer. The quality of the generated images from the quantum annealer gets worse than the ideal case of the quantum annealing and the classical Monte-Carlo sampling.
Teacher-Student Learning for a Binary Perceptron with Quantum Fluctuations Peer-reviewed

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

Journal of the Physical Society of Japan　90　(7)　074002-1-074002-11　2021/07/15
Publisher: Physical Society of Japan
DOI： 10.7566/jpsj.90.074002 　

ISSN： 0031-9015

eISSN： 1347-4073
Mean field analysis of reverse annealing for code-division multiple-access multiuser detection Peer-reviewed

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

Physical Review Research　3　(3)　033006-1-033006-22　2021/07/02
Publisher: American Physical Society (APS)
DOI： 10.1103/physrevresearch.3.033006 　

eISSN： 2643-1564
Benchmark Test of Black-box Optimization Using D-Wave Quantum Annealer Peer-reviewed

Ami S. Koshikawa, Masayuki Ohzeki, Tadashi Kadowaki, Kazuyuki Tanaka

Journal of the Physical Society of Japan　90　(6)　064001-1-064001-8　2021/06/15
Publisher: Physical Society of Japan
DOI： 10.7566/jpsj.90.064001 　

ISSN： 0031-9015

eISSN： 1347-4073
Efficient Quantum and Simulated Annealing of Potts Models Using a Half-hot Constraint Peer-reviewed

Shuntaro Okada, Masayuki Ohzeki, Kazuyuki Tanaka

Journal of the Physical Society of Japan　89　(9)　094801-1-094801-15　2020/09/15
Publisher: Japan Society of Applied Physics
DOI： 10.7566/jpsj.89.094801 　

ISSN： 0031-9015

eISSN： 1347-4073
Simple relation between frustration and transition points in diluted spin glasses Peer-reviewed

Ryoji Miyazaki, Yuta Kudo, Masayuki Ohzeki, Kazuyuki Tanaka

Physical Review E　102　(1)　012135-1-012135-7　2020/07/15
Publisher: American Physical Society (APS)
DOI： 10.1103/physreve.102.012135 　

ISSN： 2470-0045

eISSN： 2470-0053
Adaptive Thouless–Anderson–Palmer Equation for Higher-order Markov Random Fields Peer-reviewed

Chako Takahashi, Muneki Yasuda, Kazuyuki Tanaka

Journal of the Physical Society of Japan　89　(6)　064007-1-064007-9　2020/06/15
Publisher: Physical Society of Japan
DOI： 10.7566/jpsj.89.064007 　

ISSN： 0031-9015

eISSN： 1347-4073
Fair sampling by simulated annealing on quantum annealer

Masayuki Yamamoto, Masayuki Ohzeki, Kazuyuki Tanaka

Journal of the Physical Society of Japan　89　(2)　2020
Publisher: Physical Society of Japan
DOI： 10.7566/JPSJ.89.025002 　

ISSN： 1347-4073 0031-9015
Dynamics of order parameters of nonstoquastic Hamiltonians in the adaptive quantum Monte Carlo method

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

Physical Review E　99　(3)　2019/03/14
Publisher: American Physical Society
DOI： 10.1103/PhysRevE.99.032120 　

ISSN： 2470-0053 2470-0045
Difference between quantum annealing by imaginary-time and real-time Schrödinger equations of Grover’s search Peer-reviewed

Shuntaro Okada, Masayuki Ohzeki, Kazuyuki Tanaka

Journal of the Physical Society of Japan　88　(2)　024803-1-024803-6　2019/02

DOI： 10.7566/JPSJ.88.024803 　

More details Close

We confirmed the annealing time required to obtain the desired success probability for quantum annealing by the imaginary-time and real-time Schrödinger equations of Grover’s search, with two kinds of schedulings; one linearly decreases the quantum fluctuation and the other tunes the evolution rate of the Hamiltonian on the basis of the adiabatic condition. With linear scheduling, the required annealing time for quantum annealing by the imaginary-time Schrödinger equation is of order ${\rm{log } }(N)$, which is very different from ${\cal{O } }(N)$ required for that by the real-time Schrödinger equation. With the scheduling based on the adiabatic condition, the required annealing time is of order ${\sqrt{N } }$, which is identical to that by the real-time Schrödinger equation. Although the scheduling based on the adiabatic condition is optimal for the quantum annealing by the real-time Schrödinger equation, it is inefficient for that by the imaginary-time Schrödinger equation. This result implies that the optimal schedulings for the quantum annealing by the imaginary-time and real-time Schrödinger equations differ greatly, and the efficient scheduling considered with the quantum Monte Carlo method, which is based on the imaginary-time Schrödinger equation, is not necessarily effective in improving the performance of quantum annealing by the real-time Schrödinger equation. We discuss the efficient scheduling for quantum annealing by the imaginary-time Schrödinger equation with respect to the exponential decay of excited states.
Phase diagrams of one-dimensional Ising and XY models with fully connected ferromagnetic and anti-ferromagnetic quantum fluctuations Peer-reviewed

Shuntaro Okada, Masayuki Ohzeki, Kazuyuki Tanaka

Journal of the Physical Society of Japan　88　(2)　024802-1-024802-6　2019/02

DOI： 10.7566/JPSJ.88.024802 　

More details Close

We analytically investigated the phase diagrams of one-dimensional Ising and XY models. A second-order phase transition occurs in both models only with the transverse magnetic field. In this study, the fully connected ferromagnetic and anti-ferromagnetic XX interactions are applied in addition to the transverse magnetic field. The ferromagnetic XX interaction causes a first-order phase transition in both of the one-dimensional Ising and XY models. In contrast, by applying the anti-ferromagnetic one, the second-order phase transition can be avoided in the one-dimensional Ising model, and the critical amplitude becomes smaller in the one-dimensional XY model even though the second-order phase transition remains and the critical exponent does not change. Because the anti-ferromagnetic XX interaction works to destroy the order of the transverse magnetization, it begins to decrease at a larger strength of the transverse magnetic field. As a result, the susceptibility of the transverse magnetization with respect to the strength of the magnetic field becomes small. It is conjectured that the anti-ferromagnetic XX interaction is generally preferable to the ferromagnetic interaction to improve the performance of quantum annealing.
Momentum-space renormalization group transformation in Bayesian image modeling by Gaussian graphical model Peer-reviewed

Kazuyuki Tanaka, Masamichi Nakamura, Shun Kataoka, Masayuki Ohzeki, Muneki Yasuda

Journal of the Physical Society of Japan　87　(8)　085001-1-085001-2　2018/08

DOI： 10.7566/JPSJ.87.085001 　

More details Close

A new Bayesian modeling method is proposed by combining the maximization of the marginal likelihood with a momentum-space renormalization group transformation for Gaussian graphical models. Moreover, we present a scheme for computing the statistical averages of hyperparameters and mean square errors in our proposed method based on a momentum-space renormalization group transformation.
Statistical-Mechanical Analysis of Compressed Sensing for Hamiltonian Estimation of Ising Spin Glass Peer-reviewed

Chako Takahashi, Masayuki Ohzeki, Shuntaro Okada, Masayoshi Terabe, Shinichiro Taguchi, Kazuyuki Tanaka

Journal of the Physical Society of Japan　87　(7)　074001-074001　2018/07/15
Publisher: Physical Society of Japan
DOI： 10.7566/jpsj.87.074001 　

ISSN： 0031-9015

eISSN： 1347-4073

More details Close

Several powerful machines, such as the D-Wave 2000Q, dedicated to solving combinatorial optimization problems through the Ising-model formulation have been developed. To input problems into the machines, the unknown parameters of the Ising model must be determined, and this is a nontrivial task. It may be beneficial to construct a method to estimate the parameters of the Ising model from several pairs of values of the energy and spin configurations. In the present paper, we propose a simple method employing L1-norm minimization, which is based on the concept of compressed sensing. Moreover, we analyze the typical performance of our proposed method of Hamiltonian estimation by using the replica method. We also compare our analytical results through several numerical experiments using the alternating direction method of multipliers.
Linear-time algorithm in Bayesian image denoising based on gaussian markov random field Peer-reviewed

Muneki Yasuda, Junpei Watanabe, Shun Kataoka, Kazuyuki Tanaka

IEICE Transactions on Information and Systems　E101D　(6)　1629-1639　2018/06/01
Publisher: Institute of Electronics, Information and Communication, Engineers, IEICE
DOI： 10.1587/transinf.2017EDP7346 　

ISSN： 1745-1361 0916-8532
An Extension of Estimation of Critical Points in Ground State for Random Spin Systems Peer-reviewed

Masayuki Ohzeki, Yuta Kudo, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　87　(1)　015001-1-015001-2　2018/01

DOI： 10.7566/JPSJ.87.015001 　

ISSN： 0031-9015
Quantum annealing: next-generation computation and how to implement it when information is missing Peer-reviewed

Masayuki Ohzeki, Chako Takahashi, Shuntaro Okada, Masayoshi Terabe, Shinichiro Taguchi, Kazuyuki Tanaka

Nonlinear Theory and Its Applications, IEICE　9　(4)　392-405　2018
Publisher: Institute of Electronics, Information and Communications Engineers (IEICE)
DOI： 10.1587/nolta.9.392 　

eISSN： 2185-4106

More details Close

Recently, several powerful machines dedicated to solving combinatorial optimization problems through the Ising-model formulation have appeared. The trigger for the paradigm shift to a specialized machine for solving optimization problems was the D-Wave machine, which implements quantum annealing. Quantum annealing employs quantum fluctuations to find an optimal solution to an optimization problem with discrete variables. In particular, we input the optimization problem in the form of the Ising Hamiltonian, which is a specialized form of the quadratic unconstrained binary optimization problem. However, when we employ quantum annealing for a practical optimization problem, there are several issues. One typical issue is absence of the detailed form of the cost function, which characterizes the optimization problem. To input problems into specialized machines for solving an optimization problem, it is necessary to determine the unknown parameters within the Ising model. We propose a method to estimate the unknown parameters in the Ising Hamiltonian using compressed sensing. Furthermore, we analyze the theoretical limitations of our proposed method by employing the replica method, which is a sophisticated tool in statistical mechanics.
Deep neural network detects quantum phase transition Peer-reviewed

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

Journal of the Physical Society of Japan　87　(3)　033001-1-033001-4　2018
Publisher: Physical Society of Japan
DOI： 10.7566/JPSJ.87.033001 　

ISSN： 1347-4073 0031-9015
On-Demand Color Calibration for Pedestrian Tracking in Nonoverlapping Fields of View Peer-reviewed

Yuji Waizumi, Masako Omachi, Kazuyuki Tanaka

IEEE INTERNET OF THINGS JOURNAL　4　(2)　320-329　2017/04

DOI： 10.1109/JIOT.2016.2557814 　

ISSN： 2327-4662
Community Detection Algorithm Combining Stochastic Block Model and Attribute Data Clustering Peer-reviewed

Shun Kataoka, Takuto Kobayashi, Muneki Yasuda, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　85　(11)　114802-1-114802-10　2016/11

DOI： 10.7566/JPSJ.85.114802 　

ISSN： 0031-9015
Perturbative Interpretation of Adaptive Thouless-Anderson-Palmer Free Energy Peer-reviewed

Muneki Yasuda, Chako Takahashi, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　85　(7)　075001-1-075001-2　2016/07

DOI： 10.7566/JPSJ.85.075001 　

ISSN： 0031-9015
Parameter Estimation in Probabilistic Image Segmentation using Loopy Belief Propagation and Hyperprior

Tomohiro Furuichi, Shun Kataoka, Kazuyuki Tanaka

IEICE Technology Report　115　(384)　37-42　2015/12
Publisher:
ISSN： 0913-5685
A fast algorithm for the probabilistic image denoising using belief propagation and gradation conversion

Kento Yoshida, Shun Kataoka, Kazuyuki Tanaka

IEICE Technology Report　115　(384)　31-36　2015/12
The reconstruction of the missing pixels in the digital image using unmissing pixels and belief propagation

Ryuichi Mashiba, Shun Kataoka, Kazuyuki Tanaka

IEICE Technology Report　115　(384)　13-18　2015/12
Statistical analysis of loopy belief propagation in random fields Peer-reviewed

Muneki Yasuda, Shun Kataoka, Kazuyuki Tanaka

PHYSICAL REVIEW E　92　(4)　042120-1-042120-15　2015/10

DOI： 10.1103/PhysRevE.92.042120 　

ISSN： 1539-3755

eISSN： 1550-2376
Inverse Renormalization Group Transformation in Bayesian Image Segmentations Peer-reviewed

Kazuyuki Tanaka, Shun Kataoka, Muneki Yasuda, Masayuki Ohzeki

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　84　(4)　045001-1-045001-2　2015/04

DOI： 10.7566/JPSJ.84.045001 　

ISSN： 0031-9015
Bayesian reconstruction of missing observations(<Special section>Nowcast and forecast of road traffic by data fusion of various sensing data)

Kataoka Shun/Yasuda Muneki/Tanaka Kazuyuki, YASUDA Muneki, TANAKA Kazuyuki

Interdisciplinary information sciences　21　(1)　11-23　2015/03
Publisher: Tohoku University
DOI： 10.4036/iis.2015.11 　

ISSN： 1340-9050

More details Close

We focus on an interpolation method referred to Bayesian reconstruction in this paper. Whereas in standard interpolation methods missing data are interpolated deterministically, in Bayesian reconstruction, missing data are interpolated probabilistically using a Bayesian treatment. In this paper, we address the framework of Bayesian reconstruction and its application to the traffic data reconstruction problem in the field of traffic engineering. In the latter part of this paper, we describe the evaluation of the statistical performance of our Bayesian traffic reconstruction model using a statistical mechanical approach and clarify its statistical behavior.
A Study on Efficient Pedestrian Detection by Combining CNN and SVM.

Takuma Saito, Yuji Waizumi, Kazuyuki Tanaka

IEICE Technology Report　114　(521)　41-46　2015/03
24aBL-11 Statistical analysis of loopy belief propagation using replica cluster variation method

Yasuda M., Kataoka S., Tanaka K.

Meeting Abstracts of the Physical Society of Japan　70　3188-3188　2015
Publisher: The Physical Society of Japan (JPS)
DOI： 10.11316/jpsgaiyo.70.1.0_3188 　

ISSN： 2189-079X
Bayesian Image Segmentations by Potts Prior and Loopy Belief Propagation Peer-reviewed

Kazuyuki Tanaka, Shun Kataoka, Muneki Yasuda, Yuji Waizumi, Chiou-Ting Hsu

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　83　(12)　124002-1-124002-11　2014/12

DOI： 10.7566/JPSJ.83.124002 　

ISSN： 0031-9015
Traffic data reconstruction based on Markov random field modeling Peer-reviewed

Shun Kataoka, Muneki Yasuda, Cyril Furtlehner, Kazuyuki Tanaka

INVERSE PROBLEMS　30　(2)　025003-1-025003-14　2014/02

DOI： 10.1088/0266-5611/30/2/025003 　

ISSN： 0266-5611

eISSN： 1361-6420
Scene Text Recognition based on Positional Relation between Closed Curves Peer-reviewed

Yuji Waizumi, Kazuyuki Tanaka

2014 2ND INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE, MODELLING AND SIMULATION　182-186　2014

DOI： 10.1109/AIMS.2014.51 　
Character recognition using statistical evaluation of surrounding information

Naoki Uchida, Yuji Waizumi, Kazuyuki Tanaka

IEICE Technology Report　113　(346)　11-16　2013/12
Mean Field Approximation for Fields of Experts (<Special Issue>The 4th Young Scientist Meeting on Statistical Physics and Information Processing in Sendai)

Yasuda Muneki, Tanaka Kazuyuki

Interdisciplinary information sciences　19　113-119　2013/08
Publisher: Tohoku University
DOI： 10.4036/iis.2013.113 　

ISSN： 1340-9050

More details Close

Fields of experts (FoE) model, which is regarded as a higher-order Markov random field whose clique potentials are modeled by the products of experts, matches spatial structures of natural images well, and therefore, it is an efficient prior of natural images. However, the FoE model does not readily admit efficient inferences because of the complexity of landscape of its energy function. In this paper, we propose an efficient mean field approximation for the FoE model by using a perturbative expansion in statistical mechanics. Our proposed mean field approximation can be applied to the FoE under general settings and can be solved in linear time with respect to the number of pixels. In the latter part of this paper, we apply our method to the image inpainting problem, and we show it gives results being the same or better than ones given by a simple gradient method proposed in the original work.
Susceptibility propagation by using diagonal consistency Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

Physical Review E - Statistical, Nonlinear, and Soft Matter Physics　87　(1)　012134-1-012134-6　2013/01/24

DOI： 10.1103/PhysRevE.87.012134 　

ISSN： 1539-3755 1550-2376
Network Application Identification using Sequential Transition Patterns of Payload Length Peer-reviewed

Yuji Waizumi, Tsuyoshi Sato, Kazuyuki Tanaka

Interdisciplinary Information Sciences　18　(2)　189-196　2012/12
Publisher: The Editorial Committee of the Interdisciplinary Information Sciences
DOI： 10.4036/iis.2012.189 　

ISSN： 1340-9050

More details Close

A person-to-person information sharing is easily realized by P2P networks that servers are not essential. Information leakages, which are caused by malicious accesses for P2P networks, has become new social issues. To prevent information leakage, it is necessary to detect and block traffics of P2P software. Since some P2P softwares can spoof port numbers, it is difficult to detect the traffics sent from P2P softwares by using port numbers. It is more difficult to devise effective countermeasures for detecting the software because their protocol are not public. In this paper, we propose a method to identify applications using sequential transition patterns of payload length. Through real traffic experiment we show that proposed method can quickly and accurately identify network applications.
Learning Algorithm for Boltzmann Machines using Max-Product Algorithm and Pseudo-Likelihood Peer-reviewed

Muneki Yasuda, Junya Tannai, Kazuyuki Tanaka

Interdisciplinary Information Sciences　18　(1)　55-63　2012/12
Publisher: The Editorial Committee of the Interdisciplinary Information Sciences
DOI： 10.4036/iis.2012.55 　

ISSN： 1340-9050

More details Close

Boltzmann machines are parametric probabilistic models for the statistical machine learning, forming Markov random fields. Owing to their normalization constant, inference and learning in Boltzmann machines are generally classified under NP-hard problems. Maximum pseudo-likelihood estimation is an effective approximate learning method for Boltzmann machines. However, in principle, we cannot use this method for incomplete data sets, except for some special cases. In this paper, we propose a new learning algorithm for Boltzmann machines with incomplete data sets by generating a pseudo-complete data set from a given incomplete data using the max-product algorithm and the Markov chain Monte Carlo method, and then, by applying maximum pseudo-likelihood estimation to the pseudo-complete data set.
Bayesian Image Modeling by Means of a Generalized Sparse Prior and Loopy Belief Propagation Peer-reviewed

Kazuyuki Tanaka, Muneki Yasuda, D. Michael Titterington

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　81　(11)　114802-1-114802-9　2012/11

DOI： 10.1143/JPSJ.81.114802 　

ISSN： 0031-9015
Design of probabilistic image inpainting filters using Gaussian graphial models

Tomotaka Kitagawa, Muneki Yasuda, Kazuyuki Tanaka

IEICE Technology Report　112　(298)　57-62　2012/11
Composite likelihood estimation for bipartite Boltzmann machines

Gaku Asari, Muneki Yasuda, Yuji Waizumi, Kazuyuki Tanaka

IEICE Technology Report　112　(298)　39-44　2012/11
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

The recent development of information technology has enabled us to obtain and to storage huge information data. Because of this, effective usages of the information data have been one of the central points of information sciences. A Boltzmann machine is a model of the machine learning theory and can be a powerful tool for the aim. In this paper, we focus on a bipartite Boltzmann machine (BBM) consisted of two different layers: one layer is the input layer and the other is the output layer. In this model, once the input layer is clamped by given data, inference of the output layer is quite easy because of the property of conditional independence. Therefore, this model is expected to be applied to expert systems such as medical test system. However, the learning of BBM using the maximum likelihood estimation is still intractable, Because the computational complexity of learning exponentially grows with the increase in the size of system, In this paper, we propose a new learning algorithm for BBM by using the composite likelihood estimation (CLE) which is a statistical mathematical technique to approximate the MLE.
A Staudy on Web Application Identification based on Communication Characteristics

Shoki Matsuo, Yuji Waizumi, Kazuyuki Tanaka

IEICE Technical Report　112　(210)　61-66　2012/09
Replica Plefka expansion of Ising systems Peer-reviewed

Muneki Yasuda, Yoshiyuki Kabashima, Kazuyuki Tanaka

JOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT　2012　(4)　P04002-1-P04002-16　2012/04

DOI： 10.1088/1742-5468/2012/04/P04002 　

ISSN： 1742-5468
Inverse Problem in Pairwise Markov Random Fields Using Loopy Belief Propagation Peer-reviewed

Muneki Yasuda, Shun Kataoka, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　81　(4)　044801-1-044801-8　2012/04

DOI： 10.1143/JPSJ.81.044801 　

ISSN： 0031-9015
Statistical Analysis of Gaussian Image Inpainting Problems Peer-reviewed

Shun Kataoka, Muneki Yasuda, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　81　(2)　025001-1-025001-2.　2012/02

DOI： 10.1143/JPSJ.81.025001 　

ISSN： 0031-9015
Statistical analysis of the expectation-maximization algorithm with loopy belief propagation in Bayesian image modeling Peer-reviewed

Shun Kataoka, Muneki Yasuda, Kazuyuki Tanaka, D. M. Titterington

PHILOSOPHICAL MAGAZINE　92　(1-3)　50-63　2012

DOI： 10.1080/14786435.2011.624558 　

ISSN： 1478-6435
TAP equation for non-negative Boltzmann machine Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

PHILOSOPHICAL MAGAZINE　92　(1-3)　192-209　2012

DOI： 10.1080/14786435.2011.634856 　

ISSN： 1478-6435
Composite Likelihood Estimation for Restricted Boltzmann Machines Peer-reviewed

Muneki Yasuda, Shun Kataoka, Yuji Waizumi, Kazuyuki Tanaka

2012 21ST INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR 2012)　2234-2237　2012

ISSN： 1051-4651
Gaussian FoE model with Correlations among Color Components

Ryuta Murayama, Muneki Yasuda, Yuji Waizumi, Kazuyuki Tanaka

IEICE Technical Report　111　(275)　105-111　2011/11
Learning Algorithm in Restricted Boltzmann Machines using Kullback-Leibler Importance Estimation Procedure Invited Peer-reviewed

Muneki Yasuda, Tetsuharu Sakurai, Kazuyuki Tanaka

Nonlinear Theory and Its Applications, IEICE　2　(2)　153-164　2011/04
Publisher: The Institute of Electronics, Information and Communication Engineers
DOI： 10.1587/nolta/2.153 　

ISSN： 2185-4106

More details Close

Restricted Boltzmann machines (RBMs) are bipartite structured statistical neural networks and consist of two layers. One of them is a layer of visible units and the other one is a layer of hidden units. In each layer, any units do not connect to each other. RBMs have high flexibility and rich structure and have been expected to applied to various applications, for example, image and pattern recognitions, face detections and so on. However, most of computational models in RBMs are intractable and often belong to the class of NP-hard problem. In this paper, in order to construct a practical learning algorithm for them, we employ the Kullback-Leibler Importance Estimation Procedure (KLIEP) to RBMs, and give a new scheme of practical approximate learning algorithm for RBMs based on the KLIEP.
Network Application Identification Based on Communication Characteristics of Application Messages Peer-reviewed

Yuji Waizumi, Yuya Tsukabe, Hiroshi Tsunoda, Yoshiaki Nemoto, Kazuyuki Tanaka

Journal of Communication and Computer　8　(5)　111-119　2011/02
Automatic Ridge Network Detection in Crumpled Paper Based on Graph Density Peer-reviewed

Marvin Huang, Chiou-Ting Hsu, Kazuyuki Tanaka

2011 IEEE 13TH INTERNATIONAL WORKSHOP ON MULTIMEDIA SIGNAL PROCESSING (MMSP)　1-6　2011

DOI： 10.1109/MMSP.2011.6093840 　

ISSN： 2163-3517
Deterministic approximate learning algorithm for Boltzmann machines using correlation equality Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

IEICE Transactions on Information and Systems (Japanese Edition)　J93-D　(11)　2446-2453　2010/11
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 1880-4535
Advanced susceptibility propagation

Muneki Yasuda, Kazuyuki Tanaka

IEICE Technical Report　110　(265)　59-63　2010/11
Loopy belief propagation for sparse Markov random fields

Kazuyuki Tanaka

IEICE Techniocal Report　110　(295)　1-5　2010/11
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

In the present paper, new loopy belief propagations for sparse Markov random fields are proposed. Our schenmes are based on loopv belief propagations in Gaussian graphical models and adaptive TAP methods.
Network application identification using observation probability of payload length

Yuji Waizumi, Tsuyoshi Sato, Kazuyuki Tanaka

IEICE Technical Report　110　(295)　7-12　2010/11
Network application identification using sequential transition Patterns

Tsuyoshi Sato, Yuji Waizumi, Kazuyuki Tanaka

IEICE Techniocal Report　110　(170)　41-46　2010/08/06
Statistical Performance Analysis in Probabilistic Image Processing Peer-reviewed

Shun Kataoka, Muneki Yasuda, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　79　(2)　025001-1-025001-2　2010/02

DOI： 10.1143/JPSJ.79.025001 　

ISSN： 0031-9015
Statistical performance analysis by loopy belief propagation in Bayesian image modeling Peer-reviewed

Kazuyuki Tanaka, Shun Kataoka, Muneki Yasuda

INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2010 (IW-SMI 2010)　233　012013-1-012013-11　2010

DOI： 10.1088/1742-6596/233/1/012013 　

ISSN： 1742-6588
Approximate Learning Algorithm in Boltzmann Machines Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

NEURAL COMPUTATION　21　(11)　3130-3178　2009/11

DOI： 10.1162/neco.2009.08-08-844 　

ISSN： 0899-7667
Approximate Learning Algorithm in Boltzmann Machines Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

Neural Computation　21　(11)　3130-3178　2009/10/19

DOI： 10.1162/neco.2009.08-08-844 　
PROBABILISTIC IMAGE PROCESSING BY EXTENDED GAUSS-MARKOV RANDOM FIELDS Peer-reviewed

Kazuyuki Tanaka, Nicolas Morin, Muneki Yasuda, D. M. Titterington

2009 IEEE/SP 15TH WORKSHOP ON STATISTICAL SIGNAL PROCESSING, VOLS 1 AND 2　617-+　2009
Mathematical structures of loopy belief propagation and cluster variation method Peer-reviewed

Kazuyuki Tanaka

INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2008 (IW-SMI 2008)　143　012023-1-012023-18　2009

DOI： 10.1088/1742-6596/143/1/012023 　

ISSN： 1742-6588
A quantum-statistical-mechanical extension of Gaussian mixture model Peer-reviewed

Kazuyuki Tanaka, Koji Tsuda

INTERNATIONAL WORKSHOP ON STATISTICAL-MECHANICAL INFORMATICS 2007 (IW-SMI 2007)　95　012023-1-012023-9　2008

DOI： 10.1088/1742-6596/95/1/012023 　

ISSN： 1742-6588
Statistical Inferences by Gaussian Markov Random Fields on Complex Networks Peer-reviewed

Kazuyuki Tanaka, Takafumi Usui, Muneki Yasuda

2008 INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE FOR MODELLING CONTROL & AUTOMATION, VOLS 1 AND 2　214-219　2008

DOI： 10.1109/CIMCA.2008.14 　
Approximate Learning Algorithm for Restricted Boltzmann Machines Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

2008 INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE FOR MODELLING CONTROL & AUTOMATION, VOLS 1 AND 2　692-697　2008

DOI： 10.1109/CIMCA.2008.57 　
An consideration of EM algorithm with belief propagation for image restoration Peer-reviewed

Kei Inoue, Kazuyuki Tanaka

Information Technology Letters　6　187-188　2007/09/05
Statistical trajectory of an approximate EM algorithm for probabilistic image processing Peer-reviewed

Kazuyuki Tanaka, D. M. Titterington

JOURNAL OF PHYSICS A-MATHEMATICAL AND THEORETICAL　40　(37)　11285-11300　2007/09

DOI： 10.1088/1751-8113/40/37/007 　

ISSN： 1751-8113
The mathematical structure of the approximate linear response relation Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

JOURNAL OF PHYSICS A-MATHEMATICAL AND THEORETICAL　40　(33)　9993-10007　2007/08

DOI： 10.1088/1751-8113/40/33/005 　

ISSN： 1751-8113
Erratum: Replica analysis of preferential urn model (Journal of the Physical Society of Japan (2006) 75 (074802)) Peer-reviewed

J. Ohkubo, M. Yasuda, K. Tanaka

Journal of the Physical Society of Japan　76　(4)　048001-1-048001-1　2007/04
Publisher: INSTITUTE OF PURE AND APPLIED PHYSICS
DOI： 10.1143/JPSJ.76.048001 　

ISSN： 0031-9015
Boltzmann Machines with Bounded Continuous Random Variables Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

Interdisciplinary Information Sciences　13　(1)　25-31　2007/03
Publisher: Tohoku University
DOI： 10.4036/iis.2007.25 　

ISSN： 1340-9050

More details Close

We propose a Boltzmann machine formulated as a probabilistic model where every random variable takes bounded continuous values, and we derive the Thouless–Anderson–Palmer equation for the model. The proposed model includes the non-negative Boltzmann machine and the Sherrington–Kirkpatrick model with spin-S at S→∞ as a special case. It is known that the Sherrington–Kirkpatrick model with spin-S has a spin glass phase. Thus, the proposed Boltzmann machine is expected to be able to learn practical complex data.
Nonadditive volume and community detection problem in complex networks Peer-reviewed

Jun Ohkubo, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　75　(11)　115001-1-115001-2　2006/11

DOI： 10.1143/JPSJ.75.115001 　

ISSN： 0031-9015
Digital Image Inpainting Algorithm by Using Gaussian Graphical Model Peer-reviewed

Muneki Yasuda, Jun Ohkubo, Kazuyuki Tanaka

Information Technology Letters　5　225-228　2006/09/05
Publisher: Forum on Information Technology

More details Close

第5回情報科学技術フォーラム (FIT2006; September 5-7, 2006, 福岡大学) の査読付き論文
The relationship between Plefka's expansion and the cluster variation method Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　75　(8)　084006-1-084006-8　2006/08

DOI： 10.1143/JPSJ.75.084006 　

ISSN： 0031-9015
Statistical Analysis of Hyperoparameter Estimation in Probabilistic Image Processing

Hikaru Kurasawa, Kazuyuki Tanaka

IEICE Technical Report　106　(163)　7-12　2006/07/14
Replica analysis of preferential urn model Peer-reviewed

Jun Ohkubo, Muneki Yasuda, Kazuyuki Tanaka

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　75　(7)　074802-1-074802-6　2006/07

DOI： 10.1143/JPSJ.75.074802 　

ISSN： 0031-9015
Statistical learning procedure in loopy belief propagation for probabilistic image processing Peer-reviewed

Kazuyuki Tanaka

International Conference on Computational Intelligence for Modelling, Control & Automation Jointly with International Conference on Intelligent Agents, Web Technologies & Internet Commerce, Vol 2, Proceedings　II　741-746　2006
Linear response formula in generalized belief propagation for probabilistic inference Peer-reviewed

Kazuyuki Tanaka

International Conference on Computational Intelligence for Modelling, Control & Automation Jointly with International Conference on Intelligent Agents, Web Technologies & Internet Commerce, Vol 1, Proceedings　I　669-674　2006
Digital image inpainting based on Markov random field Peer-reviewed

M. Yasuda, J. Ohkubo, K. Tanaka

INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE FOR MODELLING, CONTROL & AUTOMATION JOINTLY WITH INTERNATIONAL CONFERENCE ON INTELLIGENT AGENTS, WEB TECHNOLOGIES & INTERNET COMMERCE, VOL 2, PROCEEDINGS　II　747-+　2006
Preferential urn model on a random network Peer-reviewed

Jun Ohkubo, Kazuyuki Tanaka

INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE FOR MODELLING, CONTROL & AUTOMATION JOINTLY WITH INTERNATIONAL CONFERENCE ON INTELLIGENT AGENTS, WEB TECHNOLOGIES & INTERNET COMMERCE, VOL 1, PROCEEDINGS　I　1153-+　2006
Generalized Belief Propagation Formula in Probabilistic Information Processing based on Gaussian Graphical Model Peer-reviewed

Kazuyuki Tanaka

IEICE Transactions on Information and Systems　J88-D-II　(12)　2368-2379　2005/12/01
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0915-1923
Preferential urn model and nongrowing complex networks Peer-reviewed

J Ohkubo, M Yasuda, K Tanaka

PHYSICAL REVIEW E　72　(6)　065104(R)-1-065104(R)-4　2005/12

DOI： 10.1103/PhysRevE.72.065104 　

ISSN： 1539-3755
A new update procedure for belief propagation using heterogeneity of complex networks Peer-reviewed

Jun Ohkubo, Muneki Yasuda, Kazuyuki Tanaka

Proceedings of Eighth Workshop on Information-Based Induction Sciences　179-184　2005/11/09
Estimation of the structure in Boltzman machines from learning data Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka

Proceedings of Eighth Workshop on Information-Based Induction Sciences　191-196　2005/11/09
Statistical-mechanical iterative algorithms on complex networks Peer-reviewed

J Ohkubo, M Yasuda, K Tanaka

PHYSICAL REVIEW E　72　(4)　046135-1-046135-8　2005/10

DOI： 10.1103/PhysRevE.72.046135 　

ISSN： 2470-0045

eISSN： 2470-0053
Generation of complex bipartite graphs by using a preferential rewiring process Peer-reviewed

Jun Ohkubo, Kazuyuki Tanaka, Tsuyoshi Horiguchi

Physical Review E - Statistical, Nonlinear, and Soft Matter Physics　72　(3)　2005/09

DOI： 10.1103/PhysRevE.72.036120 　

ISSN： 1539-3755 1550-2376
Generation of complex bipartite graphs by using a preferential rewiring process Peer-reviewed

J Ohkubo, K Tanaka, T Horiguchi

PHYSICAL REVIEW E　72　(3)　036120-1-036120-10　2005/09

DOI： 10.1103/PhysRevE.72.036120 　

ISSN： 2470-0045

eISSN： 2470-0053
Image Segmentation Based on Bethe Approximation for Gaussian Mixture Model Peer-reviewed

Fan Chen, Kazuyuki Tanaka, Tsuyoshi Horiguchi

Interdisciplinary Information Sciences　11　(1)　17-29　2005/03
Publisher: Tohoku University
DOI： 10.4036/iis.2005.17 　

ISSN： 1340-9050

More details Close

We propose an image segmentation algorithm under an expectation-maximum scheme using a Bethe approximation. In the stochastic image processing, the image data is usually modeled in terms of Markov random fields, which can be characterized by a Gibbs distribution. The Bethe approximation, which takes account of nearest-neighbor correlations, provides us with a better approximation to the Gibbs free energy than the commonly used mean-field approximation. We apply the Bethe approximation to the image segmentation problem and investigate its efficiency by numerical experiments. As a result, our approach shows better robustness and faster converging speed than those using the mean-field approximation.
First-order phase transition and Bayesian image processing by loopy belief propagation Peer-reviewed

K Tanaka, DM Titterington

PROGRESS OF THEORETICAL PHYSICS SUPPLEMENT　(157)　288-291　2005

ISSN： 0375-9687
Multi-scale image segmentation based on renormalization group theory Peer-reviewed

Fan Chen, Kazuyuki Tanaka, Tsuyoshi Horiguchi

Proceedings of the Seventh Workshop on Information-Based Induction Sciences　274-280　2004/11/08
Learning of Boltzmann machines without hiden unit by using cluster variation method Peer-reviewed

Muneki Yasuda, Kazuyuki Tanaka, Tsuyoshi Horiguchi

Proceeedings of the Seventh Workshop on Information-Based Induction Sciences　281-288　2004/11/08
Probabilistic information processing and belief propagation Invited

Kazuyuki Tanaka

IEICE Technical Report　104　(338)　25-32　2004/10/14
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

The basic frameworks of the Bayesian network and the belief propagation in the probabilistic information processing are reviewed in the standpoint of probabilistic inference. Some simple examples of Bayesian networks for the probabilistic inference are demonstrated and the basic formulation of the belief propagation algorithm is explained.
Bayesian network and probabilistic Inference ---Iterated algorithm for probabilistic inference based on variational approach Invited

Kazuyuki Tanaka

IEICE Technical Report　104　(348)　35-42　2004/10
Accuracy of the Bethe approximation for hyperparameter estimation in probabilistic image processing Peer-reviewed

K Tanaka, H Shouno, M Okada, DM Titterington

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL　37　(36)　8675-8695　2004/09

DOI： 10.1088/0305-4470/37/36/007 　

ISSN： 0305-4470
Statistical mechanics and information processing ---New information processing technology based on free energy Invited

Kazuyuki Tanaka

物性研究　81　(5)　679-710　2004/02/20
Dynamics and its stability of Boltzmann-machine learning algorithm for gray scale image restoration Peer-reviewed

J Inoue, K Tanaka

SLOW DYNAMICS IN COMPLEX SYSTEMS　708　731-734　2004

ISSN： 0094-243X
Belief propagation for image restoration by using Gaussian model Peer-reviewed

Kazuyuki Tanaka, Noriko Yoshiike, Hayaru Shouno, Masato Okada

Proceedings of Sixth Workshop on Information-Based Induction Sciences　83-88　2003/11
A Bayesian Approach to Human Motion Field Estimation Peer-reviewed

Noriko Yoshiike, Kazuyuki Tanaka

Proceedings of Sixth Workshop on Information-Baed Induction Sciences　65-70　2003/11
Mean field theory of EM algorithm for Bayesian grey scale image restoration Peer-reviewed

J Inoue, K Tanaka

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL　36　(43)　10997-11010　2003/10

DOI： 10.1088/0305-4470/36/43/023 　

ISSN： 0305-4470
Probabilistic image processing by means of the Bethe approximation for the Q-Ising model Peer-reviewed

K Tanaka, J Inoue, DM Titterington

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL　36　(43)　11023-11035　2003/10

DOI： 10.1088/0305-4470/36/43/025 　

ISSN： 0305-4470
Probabilistic inference by means of cluster variation method and linear response theory Peer-reviewed

K Tanaka

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS　E86D　(7)　1228-1242　2003/07

ISSN： 0916-8532
Coupled Markov random field models with phases as line field and region field Peer-reviewed

Kazuyuki Tanaka, Daiki Furusato, Tsuyoshi Horiguchi

Systems and Computers in Japan　34　(7)　24-33　2003/06/30

DOI： 10.1002/scj.10364 　

ISSN： 0882-1666
Color Image Compression Algorithm Using Self-Organizing Feature Map Peer-reviewed

Kazuyuki Tanaka, Norihiro Hoshi, Tsuyoshi Horiguchi

Interdisplinary Information Sciences　9　(2)　201-208　2003/01
Publisher: Tohoku University
DOI： 10.4036/iis.2003.201 　

ISSN： 1340-9050

More details Close

A new color image compression algorithm using Kohonen's self-organizing feature map is proposed. Our algorithm is an extension of color image compression algorithm proposed by Pei and Lo [IEEE Trans. Circuits Syst. Video Technol., 8: 191–205 (1998)]. N neurons are introduced in order to reduce a given full color image with 2²⁴ colors to an indexed color image with N colors. There are control parameters for the competitive learning among neurons in the self-organizing feature map algorithm. In our proposed algorithm, some of the control parameters, which are included in a neighboring function defined for neurons, are updated by taking relationship among neighboring neurons into account, though all control parameters are updated so as to decrease monotonically and exponentially with respect to each iteration step in Pei and Lo's algorithm. The color palette obtained by the proposed algorithm is more robust as for control parameters than that by Pei and Lo's algorithm.
Loopy belief propagation and probabilistic image processing Peer-reviewed

K Tanaka, JI Inoue, DM Titterington

2003 IEEE XIII WORKSHOP ON NEURAL NETWORKS FOR SIGNAL PROCESSING - NNSP'03　13　329-338　2003

DOI： 10.1109/NNSP.2003.1318032 　
Statistical-mechanical approach to image processing Invited Peer-reviewed

Kazuyuki Tanaka

Journal of Physics A: Mathematical and General　35　(37)　R81-R150　2002/09/20

DOI： 10.1088/0305-4470/35/37/201 　

ISSN： 0305-4470
Coupled Markov Random Field Models with Phases as Line Field and Region Field Peer-reviewed

Kazuyuki Tanaka, Daiki Furusato, Tsuyoshi Horiguchi

The Transactions of the Institute of Electronics, Information and Communication Engineering D-II　J85-D-II　(5)　805-814　2002/05
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0915-1923
Automatic Hyperparameter Estimation in Probabilistic Image Restoration Algorithm Based on Gibbs Microcanonical Distribution Peer-reviewed

Kazuyuki Tanaka

The Transactions of the Institute of Electronics, Information and Communication Engineering D-II　J85-D-II　(5)　815-824　2002/05
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0915-1923
Solvable Markov random field model in color image restoration Peer-reviewed

K Tanaka, T Horiguchi

PHYSICAL REVIEW E　65　(4)　046124-1-046124-18　2002/04

DOI： 10.1103/PhysRevE.65.046142 　

ISSN： 1539-3755
Maximum likelihood hyperparameter estimation for solvable Markov random field model in image restoration Peer-reviewed

K Tanaka, J Inoue

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS　E85D　(3)　546-557　2002/03

ISSN： 0916-8532
Solvable Markov random field model in color image restoration Peer-reviewed

Kazuyuki Tanaka, Tsuyoshi Horiguchi

Physical Review E - Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics　65　(4)　18　2002

DOI： 10.1103/PhysRevE.65.046142 　

ISSN： 1063-651X
Dynamics of the maximum marginal likelihood hyperparameter estimation in image restoration: Gradient descent versus expectation and maximization algorithm Peer-reviewed

J Inoue, K Tanaka

PHYSICAL REVIEW E　65　(1)　016125-1-016125-11　2002/01

DOI： 10.1103/PhysRevE.65.016125 　

ISSN： 1539-3755
Coupled Markov Random Field Models with Line Field and Region Field based on Plane Rotator Model Peer-reviewed

Kazuyuki Tanaka, Daiki Furusato, Tsuyoshi Horiguchi

Proceedings of Fourth Workshop on Information-Based Induction Sciences　255-260　2001/07
A Formulation of Gray-level Image Restoration based on Constrained Optimization Peer-reviewed

Kazuyuki Tanaka

Proceedings of Fourth Workshop on Information-based Induction Sciences　159-164　2001/07
Image Restorations by using Compound Gauss-Markov Random Field Model with Quantumized Line Fields Peer-reviewed

Kazuyuki Tanaka

Transactions of the Institute of Electronics, Information and Communication Engineerings D-II　J84-D-II　(4)　737-743　2001/04
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0915-1923
Maximum marginal likelihood estimation and constrained optimization in image restoration Peer-reviewed

Kazuyuki Tanaka

Transactions of the Japanese Society for Artificial Intelligence　16　(2)　246-258　2001

DOI： 10.1527/tjsai.16.246 　

ISSN： 1346-0714 1346-8030
Staistical-mechanical iterative algorithm by means of cluster variation method in compound Gauss-Markov random field model Peer-reviewed

Kazuyuki Tanaka

Transactions of the Japanese Society for Artificial Intelligence　16　(2)　259-267　2001

DOI： 10.1527/tjsai.16.259 　

ISSN： 1346-0714 1346-8030
Dynamical properties of image restoration and hyper-parameter estimation Peer-reviewed

J Inoue, K Tanaka

NEURAL NETWORKS FOR SIGNAL PROCESSING XI　11　383-392　2001
Theoretical Study of Hyperparameter Estimation by Maximization of Marginal Likelihood in Image Restoration by means of Cluster Variation Method Peer-reviewed

Kazuyuki Tanaka

Transactions of the Institute of Electronics, Information and Communication Engineerings A　J83-A　(10)　1148-1160　2000/10
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5707
Statistical-mechanical Iterative Algorithms for Image Restorations by using Compound Gauss-Markov Random Field Model Peer-reviewed

Kazuyuki Tanaka

Proceedings of Third Workshop on Information-Based Induction Sciences　39-44　2000/07
Probabilistic Model and Image Processing Peer-reviewed

Kazuyuki Tanaka

Proceedings of Third Workshop on Information-Based Induction Sciences　283-288　2000/07
Quantum statistical-mechanical iterative method in image restoration Peer-reviewed

Kazuyuki Tanaka, Tsuyoshi Horiguchi

Electronics and Communications in Japan, Part III: Fundamental Electronic Science (English translation of Denshi Tsushin Gakkai Ronbunshi)　83　(3)　84-94　2000
Publisher: Scripta Technica Inc
DOI： 10.1002/(SICI)1520-6440(200003)83:3<84::AID-ECJC9>3.0.CO;2-N 　

ISSN： 1042-0967
Probabilistic computational method in image restoration based on statistical-mechanical technique Peer-reviewed

K Tanaka

SOFT COMPUTING IN INDUSTRIAL APPLICATIONS　401-414　2000
Maximum Posterior Marginal Estimate in Statistical Method for Image Restoration Peer-reviewed

Kazuyuki Tanaka

Transactions of the Institute of Electronics,Information and Communication Engineerings A　J82-A　(10)　1679-1684　1999/10
Approximate Optimization Algorithms in Markov Random Field Model based on Statistical-mechanical Techniques Peer-reviewed

Kazuyuki Tanaka

Proceedings of International Workshop on Soft Computing in Industry'99　165-170　1999/07
Cluster ZTP in the recovery of an image Peer-reviewed

T Morita, K Tanaka

PATTERN RECOGNITION LETTERS　18　(14)　1479-1493　1997/12

DOI： 10.1016/S0167-8655(97)00146-3 　

ISSN： 0167-8655
Quantum Statistical-Mechanical Iterative Method in Image Restoration Peer-reviewed

Kazuyuki Tanaka, Tsuyoshi Horiguchi

The Transactions of the Institute of Electronics, Information and Communication Engineering A　J80-A　(12)　2117-2126　1997/12
濃淡画像の画像修復に対する統計力学的反復計算アルゴリズム Peer-reviewed

Kazuyuki Tanaka, Tohru Morita

The Transactions of the Institute of Electoronics, Information and Communication Engineering A　J80-A　(6)　1033-1037　1997/06
画像修復に対する統計力学的反復計算法 Peer-reviewed

Kazuyuki Tanaka, Mikako Ichioka, Tohru Morita

The Transactions of thee Electronics, Information and Communication Engineering A　J80-A　(1)　260-268　1997/01
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5707
Statistical-mechanical Study of the Image Restoration based on the Markov Random Field Model Peer-reviewed

Kazuyuki Tanaka, Tohru Morita

Proceedings of the International Conference on Soft Computing in Industory　78-83　1996/04
Determination of parameters in an image recovery by statistical-mechanical means Peer-reviewed

T Morita, K Tanaka

PHYSICA A　223　(1-2)　244-262　1996/01

DOI： 10.1016/0378-4371(95)00276-6 　

ISSN： 0378-4371
Correlated Random Field Ising Model and Image Restoration Problem Peer-reviewed

Kazuyuki Tanaka, Tohru Morita

Proceedings of International Symposium on Coferent Approaches to Fluctuations　150-153　1996
Statistical-mechanical algorithm in MRF model based on variational principle Peer-reviewed

Kazuyuki Tanaka, Mikako Ichioka, Tohru Morita

Proceedings - International Conference on Pattern Recognition　2　381-388　1996
Publisher: Institute of Electrical and Electronics Engineers Inc.
DOI： 10.1109/ICPR.1996.546853 　

ISSN： 1051-4651
CLUSTER VARIATION METHOD AND IMAGE-RESTORATION PROBLEM Peer-reviewed

K TANAKA, T MORITA

PHYSICS LETTERS A　203　(2-3)　122-128　1995/07

DOI： 10.1016/0375-9601(95)00387-I 　

ISSN： 0375-9601
NUMERICAL STUDY OF 2D ANNNI MODEL IN THE CLUSTER VARIATION METHOD Peer-reviewed

Y MURAI, K TANAKA, T MORITA

PHYSICA A　217　(1-2)　214-226　1995/07

DOI： 10.1016/0378-4371(95)00075-I 　

ISSN： 0378-4371
CRITICAL-BEHAVIOR OF A FRUSTRATED SPIN-1 ISING-MODEL Peer-reviewed

Y HONDA, K TANAKA, T HORIGUCHI

PHYSICAL REVIEW B　51　(21)　15260-15266　1995/06

DOI： 10.1103/PhysRevB.51.15260 　

ISSN： 0163-1829
A LATTICE MODEL OF FLUX LINES Peer-reviewed

T MORITA, K TANAKA

PHYSICS LETTERS A　202　(1)　133-136　1995/06

DOI： 10.1016/0375-9601(95)00259-6 　

ISSN： 0375-9601
APPLICATION OF THE CLUSTER VARIATION METHOD TO 3-DIMENSIONAL T-J-MODEL-II Peer-reviewed

K TANAKA, T MORITA

JOURNAL OF THE KOREAN PHYSICAL SOCIETY　28　S373-S377　1995/05

ISSN： 0374-4884

eISSN： 1976-8524
GENERALIZED HEISENBERG-MODEL OF SPIN-S-GREATER-THAN-1/2 EQUIVALENT-TO THOSE OF SPIN-1/2 OR SPIN-1 Peer-reviewed

T MORITA, K TANAKA

JOURNAL OF THE KOREAN PHYSICAL SOCIETY　28　S378-S383　1995/05

ISSN： 0374-4884

eISSN： 1976-8524
CLUSTER VARIATION METHOD FOR 2-DIMENSIONAL HUBBARD AND T-J MODELS Peer-reviewed

K TANAKA, H EBISAWA, T MORITA

PHYSICA C-SUPERCONDUCTIVITY AND ITS APPLICATIONS　235　(PT.4)　2187-2188　1994/12

DOI： 10.1016/0921-4534(94)92315-9 　

ISSN： 0921-4534
APPLICATION OF CLUSTER VARIATION METHOD TO 3-DIMENSIONAL T-J MODEL Peer-reviewed

K TANAKA, Y TAKANE, H EBISAWA, T MORITA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　63　(8)　2909-2918　1994/08

DOI： 10.1143/JPSJ.63.2909 　

ISSN： 0031-9015
GENERALIZED T-J MODEL WITH GENERAL HALF-ODD-INTEGRAL SPIN-S Peer-reviewed

T MORITA, K TANAKA

PHYSICS LETTERS A　188　(1)　101-105　1994/05

DOI： 10.1016/0375-9601(94)90125-2 　

ISSN： 0375-9601
3-DIMENSIONAL T-J MODEL AND CLUSTER VARIATION METHOD Peer-reviewed

K TANAKA, Y TAKANE, H EBISAWA, T MORITA

PHYSICA B　194　481-482　1994/02

ISSN： 0921-4526
COHERENT-ANOMALY ANALYSIS WITH CLUSTER VARIATION METHOD FOR 2-DIMENSIONAL ISING-MODEL WITH NEAREST-NEIGHBOR AND NEXT-NEAREST-NEIGHBOR INTERACTIONS Peer-reviewed

K TANAKA, T HORIGUCHI, T MORITA

PROGRESS OF THEORETICAL PHYSICS SUPPLEMENT　(115)　221-235　1994

ISSN： 0375-9687
GENERALIZED HEISENBERG-MODEL OF SPIN-S GREATER-THAN 1/2 EQUIVALENT TO THAT OF SPIN 1/2 Peer-reviewed

T MORITA, K TANAKA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　62　(11)　3771-3773　1993/11

DOI： 10.1143/JPSJ.62.3771 　

ISSN： 0031-9015
ASYMPTOTIC-BEHAVIOR OF SPIN-PAIR CORRELATION-FUNCTION OF ISING-MODEL ON GENERALIZED KAGOME LATTICE Peer-reviewed

K TANAKA, T MORITA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　62　(3)　868-872　1993/03

DOI： 10.1143/JPSJ.62.868 　

ISSN： 0031-9015
DIAGRAMMATICAL TECHNIQUES FOR 2-DIMENSIONAL ISING-MODELS .3. ISING-MODEL TO VERTEX MODEL Peer-reviewed

T MORITA, K TANAKA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　62　(3)　873-879　1993/03

DOI： 10.1143/JPSJ.62.873 　

ISSN： 0031-9015
CRITICAL INDEXES FOR THE 2-DIMENSIONAL ISING-MODEL WITH NEAREST-NEIGHBOR AND NEXT-NEAREST-NEIGHBOR INTERACTIONS .2. STRIP CLUSTER APPROXIMATION Peer-reviewed

K TANAKA, T HORIGUCHI, T MORITA

PHYSICA A　192　(4)　647-664　1993/02

DOI： 10.1016/0378-4371(93)90114-J 　

ISSN： 0378-4371
ASYMPTOTIC-BEHAVIOR OF SPIN-PAIR CORRELATION-FUNCTION OF ISING-MODEL ON CHECKERBOARD LATTICE Peer-reviewed

K TANAKA, T MORITA

PROGRESS OF THEORETICAL PHYSICS　88　(5)　865-880　1992/11

DOI： 10.1143/PTP.88.865 　

ISSN： 0033-068X
CRITICAL INDEXES FOR THE 2-DIMENSIONAL ISING-MODEL WITH NEAREST-NEIGHBOR AND NEXT-NEAREST-NEIGHBOR INTERACTIONS Peer-reviewed

K TANAKA, T HORIGUCHI, T MORITA

PHYSICS LETTERS A　165　(3)　266-270　1992/05

DOI： 10.1016/0375-9601(92)90048-Q 　

ISSN： 0375-9601

eISSN： 1873-2429
LOW-TEMPERATURE BEHAVIOR OF ANTIFERROMAGNETIC ISING-MODEL ON TRIANGULAR LATTICE Peer-reviewed

T HORIGUCHI, K TANAKA, T MORITA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　61　(1)　64-69　1992/01

DOI： 10.1143/JPSJ.61.64 　

ISSN： 0031-9015
FREE-ENERGY FOR LAYERED FREE FERMION MODELS Peer-reviewed

K TANAKA, T MORITA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　61　(1)　92-101　1992/01

DOI： 10.1143/JPSJ.61.92 　

ISSN： 0031-9015
COHERENT-ANOMALY ANALYSIS WITH CLUSTER VARIATION METHOD FOR SPIN-PAIR CORRELATION-FUNCTION OF ISING-MODEL ON SQUARE LATTICE Peer-reviewed

K TANAKA, T HORIGUCHI, T MORITA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　60　(8)　2576-2587　1991/08

DOI： 10.1143/JPSJ.60.2576 　

ISSN： 0031-9015
SPIN PAIR CORRELATION-FUNCTION OF THE ISING-MODEL ON THE BRICKWORK LATTICE WITH 2ND NEIGHBOR INTERACTIONS Peer-reviewed

K TANAKA, T MORITA, K HIROIKE

PHYSICA A　171　(2)　350-373　1991/02

DOI： 10.1016/0378-4371(91)90283-I 　

ISSN： 0378-4371
COHERENT-ANOMALY METHOD FOR THE WAVE-NUMBER DEPENDENCE OF THE SUSCEPTIBILITY Peer-reviewed

T HORIGUCHI, K TANAKA, T MORITA

JOURNAL OF THE PHYSICAL SOCIETY OF JAPAN　59　(11)　4196-4197　1990/11

DOI： 10.1143/JPSJ.59.4196 　

ISSN： 0031-9015
ASYMPTOTIC BEHAVIORS OF MODIFIED BLOCK TOEPLITZ DETERMINANT Peer-reviewed

K TANAKA, T MORITA

PROGRESS OF THEORETICAL PHYSICS　84　(3)　392-409　1990/09

DOI： 10.1143/PTP.84.392 　

ISSN： 0033-068X
ASYMPTOTIC FORMS OF SOME BLOCK TOEPLITZ DETERMINANTS Peer-reviewed

K TANAKA, T MORITA

PROGRESS OF THEORETICAL PHYSICS　84　(3)　410-414　1990/09

DOI： 10.1143/PTP.84.410 　

ISSN： 0033-068X
ONE-DIMENSIONAL BINARY MIXTURE OF HARD-RODS WITH NONADDITIVE DISTANCES OF CONTACT Peer-reviewed

T MORITA, K TANAKA, K HIROIKE

MOLECULAR PHYSICS　61　(6)　1549-1561　1987/08

DOI： 10.1080/00268978700101981 　

ISSN： 0026-8976
CORRELATION-FUNCTIONS OF A ONE-DIMENSIONAL LATTICE GAS WITH NEAREST-NEIGHBOR AND NEXT-NEAREST-NEIGHBOR INTERACTIONS Peer-reviewed

K TANAKA, T MORITA, K HIROIKE

PROGRESS OF THEORETICAL PHYSICS　77　(1)　68-83　1987/01

DOI： 10.1143/PTP.77.68 　

ISSN： 0033-068X
CORRELATION-FUNCTIONS OF A ONE-DIMENSIONAL LATTICE GAS WITH INTERACTIONS UP TO 3RD-NEAREST NEIGHBORS Peer-reviewed

K TANAKA, T MORITA, K HIROIKE

PROGRESS OF THEORETICAL PHYSICS SUPPLEMENT　(87)　214-228　1986

ISSN： 0375-9687

Show all ︎Show first 5

Misc. 134

A Study of Bases Definition in Sparse Representation based Classification for Face Recognition

116　(259)　43-48　2016/10/20
Publisher: 電子情報通信学会
ISSN： 0913-5685
8aAY-1 Background Segmentation using Markov Random Field

Kataoka S., Tanaka K.

Meeting abstracts of the Physical Society of Japan　69　(2)　130-130　2014/08/22
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
A Web Application Identification based on Network Traffic Analysis

WAIZUMI Yuji, TANAKA Kazuyuki

IEICE technical report. Communication systems　113　(207)　61-66　2013/09/12
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

Recently, websites which offer various Web Applications, such as Web Mail, Video Sharing, and so on, has been increasing. In terms of network management, these applications from an identical web site should be classified into permitted and forbidden. Since these applications cannot be identified by using port number or IP address, we need a method which can identify Web Applications without using port number or IP address. In this paper, we suppose that characteristics of Web Application come in a sequence of packet length, and survey the identification performance based on it through experiments with real traffic.
Image Segmentation using Conditional Random Field

Kataoka S., Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　68　(2)　252-252　2013/08/26
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Probabilistic Graphical Model ---Bayesian Network and Related Topics

Muneki Yasuda, Shun Kataoka, Kazuyuki Tanaka

Operations Research　58　(4)　191-197　2013/04
Publisher: The Operations Research Society of Japan
ISSN： 0030-3674
27aXZD-9 Mean field approximation of higher-order Markov random fields for image processing

Yasud Muneki, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　68　(1)　349-349　2013/03/26
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
複数の領域評価手法の統合による文字認識

佐藤沢彦, 和泉勇治, 田中和之

第75回全国大会講演論文集　2013　(1)　451-452　2013/03/06

More details Close

情景画像中の文字認識においては，文字と風景の境界が正しく領域分割できなければ，認識が行えないという問題がある．これの解決策として，複数の領域分割基準による分割の結果を統合する方法が想定される．このような方法においては，分割結果統合のためのパラメータが新たに生じることになると考えられる．本研究では新たに生じたパラメータを自動設定し，複数の分割結果を効果的に統合する手法を提案する．
18pAA-2 Bayesian Modeling by Belief Propagation based on Prior with Sparse Interactions

Tanaka K., Yasuda M.

Meeting abstracts of the Physical Society of Japan　67　(2)　228-228　2012/08/24
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
18pAA-1 Plefka expansion for Markov random fields using orthogonal function expansion

Yasud Muneki, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　67　(2)　228-228　2012/08/24
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
18aAA-10 Markov Network Modeling for Traffic Density Prediction

Kataoka S., Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　67　(2)　222-222　2012/08/24
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Item Response Theory with Correlations among Items

YASUDA Muneki, TANAKA Kazuyuki

IEICE technical report. Neurocomputing　111　(483)　387-391　2012/03/07
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

Item response theory (IRT) is a recent test theory which has been developed in mainly psychology and social science. A main object of IRT is to estimate both the ability of each examinee and the difficulty of each test item, and subsequently, to provide the score of each examinee using the estimates. The conventional models in IRT cannot consider interactions among test items because they make assumption of the statistical independence among test items. In this paper, we propose a new model of IRT which can consider interactions among test items by merging IRT with the theory of Markov random field.
27pAG-6 Computational Model for Image Inpainting based on Compressed Sensing

Kataoka S., Yasuda M., Kabashima Y., Tanaka K.

Meeting abstracts of the Physical Society of Japan　67　(1)　377-377　2012/03/05
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
22aGA-7 Image compression using randomly missing data

Kataoka S., Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　66　(2)　218-218　2011/08/24
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
21pGU-8 Advanced Susceptibility Propagation and Its a Generalization

Yasuda Muneki, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　66　(2)　213-213　2011/08/24
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
28aTD-11 Advanced Susceptibility Propagation

Yasud Muneki, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　66　(1)　343-343　2011/03/03
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Special section on developments of vision computing based on probabilistic information processing: Editorial preface

Kazuyuki Tanaka

The Institute of Electronics, Information and Communication Engineerings　93　(9)　743-743　2010/09/01
Publisher: The Journal of the Institute of Electronics, Information and Communication Engineerings
ISSN： 0913-5693
Deepening and expansion of statistical inference and learning theory in image processing

Kazuyuki Tanaka, Shun Kataoka, Muneki Yasuda

The Journal of the Institute of Electronics, Information and Communication Engineerings　93　(9)　744-748　2010/09/01
Publisher: The Institute of Electronics, Information and Communication Engineerings
ISSN： 0913-5693
25pTD-7 Susceptibility Propagation using 1-RSB version of TAP equation

Yasuda Muneki, Kabashima Yoshiyuki, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　65　(2)　282-282　2010/08/18
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
20pEA-1 Re-interpretation of the adaptive TAP equation as a perturbative approximation

Yasuda Muneki, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　65　(1)　273-273　2010/03/01
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
20aEA-3 Statistical analysis of EM algorithm for probabilistic image processing by belief propagation

Kataoka S., Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　65　(1)　263-263　2010/03/01
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
25aQL-2 Statistical performance evaluation of the image restoration by using the Belief Propagation

Kataoka S., Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　64　(2)　156-156　2009/08/18
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
25aQL-5 Approximate Learning Algorithm using Replicated Adaptive TAP Equation

Yasuda M, Kabashima Y, Tanaka K

Meeting abstracts of the Physical Society of Japan　64　(2)　157-157　2009/08/18
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Probabilistic Digital Image Filters

YASUDA Muneki, TANAKA Kazuyuki

Systems, control and information　53　(8)　360-361　2009/08/15
Publisher: Institute of Systems, Control and Information Engineers
ISSN： 0916-1600
Probabilistic Image Processing Filters

Muneki Yasuda, Kazuyuki Tanaka

Systems, Control and Information　53　(8)　56-57　2009/08
Publisher: The Institute of Systems, Controll and Information Engineerings
Deeping and Expansion of Large-Scale Random Fields and Probabilistic Image Processing

TANAKA KAZUYUKI

2009　(36)　1-9　2009/06/02
Publisher: 情報処理学会
ISSN： 0919-6072
30pTF-5 Extensions of Generalized Belief Propagation by Cluster Variation Method

Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　64　(1)　346-346　2009/03/03
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
30pTF-8 Application for the statistical learning theory using the replicated adaptive TAP equation

Yasuda M, Kabashima Y, Tanaka K

Meeting abstracts of the Physical Society of Japan　64　(1)　347-347　2009/03/03
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
ビリーフプロパゲーション法を用いたボルツマンマシンの学習アルゴリズム

丹内隼也, 安田宗樹, 田中和之

電子情報通信学会技術研究報告　108　(480)　243-248　2009
Learning algorithm in Boltzmann machines using the belief propagation algorithm

丹内隼也, 安田宗樹, 田中和之

IEICE Technical Report　108　(480)　243-248　2009
22pVC-9 Mean-Field Theory for Boltzmann Machines with Hidden Units

Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　63　(2)　249-249　2008/08/25
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
22pVC-12 Probabilistic Model for Cycles in Planar Graph

Tanaka Kazuyuki, Kerrebroeck Valery Van

Meeting abstracts of the Physical Society of Japan　63　(2)　250-250　2008/08/25
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Learning and Inference of Probabilistic Causal Structure Model for Indoor Behavior Prediction of Infants

KAWATA Satoshi, MOTOMURA Yoichi, NISHIDA Yoshifumi, ISHIKAWA Shouzou, TANAKA Kazuyuki

IEICE technical report　107　(542)　279-282　2008/03/05
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

In order to contribute to saving children exposed to daily dangers, monitoring children's everyday behavior is essential for the paradigms of infant risk prevention. In this research, we introduce a statistical lerning to predict infant behavioral models. With the rapid progress in computer-aided processing, we have come to make use of precise graphical models based on calculation using Bayesian networks. Utilizing diverse graphical models will lead us to cover various prior knowledge dwelling in the real world. In this paper, we model probabilistic causal structure chronologically making use of diverse dynamic images and ultrasound location sensor data based on Bayesian Network. Compared with the maximum likelihood estimator dealing with machine learning, this model will give us broader perspective to predict expected human behavior more precisely in daily life.
スケールフリーネットワーク上のガウシアングラフィカルモデル

臼井崇文, 安田宗樹, 田中和之

電子情報通信学会技術研究報告　107　(542)　1-6　2008
Gaussian Graphical Model on Scale-Free Network

Tanafumi Usui, Muneki Yasuda, Kazuyuki Tanaka

IEICE Technical Report　107　(542)　1-6　2008
確率伝搬法を用いたEMアルゴリズムによる画像処理

井上佳, 安田宗樹, 田中和之

IEICE Technical Report　107　(542)　37-42　2008
21aTL-12 Gaussian Graphical Model on the Scale-Free Network

Usui Takafumi, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　62　(2)　256-256　2007/08/21
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
21aTL-13 Mean Field Theory for Item Response Theory with Correlations

Tanaka Kazuyuki, Sugaya Fumiaki

Meeting abstracts of the Physical Society of Japan　62　(2)　256-256　2007/08/21
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
21aTL-6 Boltzmann Machines with Bounded Continuous Random Variables

Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　62　(2)　255-255　2007/08/21
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
21pWA-5 Advanced Mean-Field Method in Probabilistic Inference for Quantum Systems

Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　62　(1)　318-318　2007/02/28
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
画像処理における確率伝搬法とEMアルゴリズムの統計的性能評価(情報物理学の数学的構造)

田中和之

数理解析研究所講究録　1532　60-69　2007/02
Publisher: 京都大学
ISSN： 1880-2818
A_019 Community Detection Algorithm by Using Entropy of Mixing

Ohkubo Jun, Tanaka Kazuyuki

5　(1)　39-40　2006/08/21
Publisher: Forum on Information Technology
26aXD-8 Statistical Analysis of EM Algorithm with Belief Propagation for Image Processing

Tanaka Kazuyuki, Minagawa Marika

Meeting abstracts of the Physical Society of Japan　61　(2)　227-227　2006/08/18
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
26aXD-9 Application of EM algorithm based on MCMC method to image restoration and its statistical performance

Inoue J., Kurasawa H., Tanaka K.

Meeting abstracts of the Physical Society of Japan　61　(2)　227-227　2006/08/18
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
26aXE-3 Urn models with network structures and power low

Ohkubo J., Tanaka K.

Meeting abstracts of the Physical Society of Japan　61　(2)　228-228　2006/08/18
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
統計力学の基礎 ---複雑ネットワークとの関連に基づいて--- (特集/ネットワーク科学の数理)

大久保潤, 田中和之

Mathematical Sciences　44　(8)　24-29　2006/08
Publisher: サイエンス社
ISSN： 0386-2240
トレンドキーワード（第2回）「スケールフリーネットワーク」/大久保潤氏・田中和之氏（東北大学大学院）の「スケールフリーネットワーク」

大久保潤, 田中和之

情報・システムソサイエティ誌　11　(1)　18-19　2006/06/01
30pUA-10 Statistical analysis of hyperparameter estimation in image restoration using a generating function method

Kurasawa Hikaru, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　61　(1)　336-336　2006/03/04
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
30aRD-6 Analysis of an urn model with randomness

Ohkubo J., Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　61　(1)　324-324　2006/03/04
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
30pUA-12 Correlations based on Plefka expansion and the linear response theory

Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　61　(1)　336-336　2006/03/04
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Probabilistic Inference and Estimation in Large-scale Random Fields

Kazuyuki Tanaka

The Journal of the Institute of Electronics, Information and Communication Engineerings　88　(9)　698-702　2005/09/01
Publisher: The Institute of Electronics, Information and Communication Engineerings
Intriduction: New Research Paradaigm of Probabilistic Inference and Statistical Learning

Kazuyuki Tanaka, Yoshiyuki Kabashima, Toshiyuki Tanaka

The Journal of the Instutitute of Electronics, Information and Communication Engineerings　88　(9)　696-697　2005/09/01
Publisher: The Institute of Electronics, Information and Comunication Engineerings
Mathematics of Bayesian Network and Probabilistic Inference

Kazuyuki Tanaka

Mathematical Sciences　43　(9)　77-83　2005/09/01
Publisher: サイエンス社
ISSN： 0386-2240
19pYO-7 Interpretation of Plefka's expansion by means of a cluster variation method

Yasuda M., Tanaka K.

Meeting abstracts of the Physical Society of Japan　60　(2)　157-157　2005/08/19
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
19pYO-10 Non-growing models for complex networks by using nonlinear Differential rewiring processes

Ohkubo J., Tanaka K.

Meeting abstracts of the Physical Society of Japan　60　(2)　158-158　2005/08/19
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
領域11「確率的情報処理の生み出す平均場理論の新展開」(2004年秋季大会シンポジウム(物性領域)の報告)

田中和之

日本物理學會誌　60　(2)　138-138　2005/02/05
Publisher: 一般社団法人日本物理学会
ISSN： 0029-0181
Statistical-Mechanical Iterative Algorithms on Complex Networks

Jun Ohkubo, Muneki Yasuda, Kazuyuki Tanaka

Physical Review E　72　(4,Pt.2)　046135　2005

DOI： 10.1103/PhysRevE.72.046135 　

ISSN： 1539-3755
Probabilistic Information Processing ---New Trend on Information Processing by Probabilistic Model and Statistical-Mechanical Method

Kazuyuki Tanaka, Yoshiyuki Kabashima, Toshiyuki Tanaka

Mathematical Sciences　42　(11)　61-68　2004/11/01
Publisher: SAIENSU-SHA Co.,Ltd
14aTD-5 Statistical-mechanical study for image restration by using belief propagation

Hayashi Y, Tanaka K, Horiguchi T

Meeting abstracts of the Physical Society of Japan　59　(2)　235-235　2004/08/25
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
14pTD-9 Learning of Boltzmann Machines by Using Cluster Variation Method

Yasuda M, Tanaka K, Horiguchi T

Meeting abstracts of the Physical Society of Japan　59　(2)　270-270　2004/08/25
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
12pTB-1 はじめに平均場理論を用いた情報処理の最近の動向(主題 : 確率的情報処理の生み出す平均場理論の新展開, 領域 11)

田中和之

日本物理学会講演概要集　59　(2)　197-197　2004/08/25
Publisher: 一般社団法人日本物理学会
ISSN： 1342-8349
14aTD-6 Generalized Belief Propagation for Gaussian graphical model

Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　59　(2)　235-235　2004/08/25
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
30aWB-6 Correctness of Bethe approximation in probabilistic image processing by means of Gaussian graphical model

Tanaka Kazuyuki, Shouno Hayau, Okada Masato

Meeting abstracts of the Physical Society of Japan　59　(1)　337-337　2004/03/03
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Probabilistic Models and Image Analysis ---Novel Computational Paradigm based on Fluctuation

Kazuyuki Tanaka

Mathematical Sciences　42　(3)　15-21　2004/03/01
Publisher: SAIENSU-SHA Co.,Ltd
リレｰ連載/確率的情報処理と統計力学 -様々なアプロｰチとそのチュｰトリアル-

田中和之

数理科学　2004
[チュｰトリアル講演] 確率的情報処理と確率伝搬型の基礎

田中和之

信学技法　2004
Publisher: コンピュｰテｰション研究会
Image compression and decompression by using Gaussian mixture model

Ito Y., Tanaka K., Horiguchi T.

Meeting abstracts of the Physical Society of Japan　58　(2)　227-227　2003/08/15
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Large-Scale Statistical Model and Information Processing ---Science of Material and Art of Information Processing

Kazuyuki Tanaka, Yoshiyuki Kabashima, Toshiyuki Tanaka

Journal of the Society of Instrument and Control Engineerings　42　(8)　622-625　2003/08
Publisher: 社団法人計測自動制御学会
DOI： 10.11499/sicejl1962.42.622 　

ISSN： 0453-4662
Probabilistic Image Processing Algorithm by means of Statistical Mechanical Method ---Belief Propagation and Statistical Mechanics

Kazuyuki Tanaka

Journal of the Society of Instrument and Control Engineerings　42　(8)　631-636　2003/08
Publisher: 社団法人計測自動制御学会
DOI： 10.11499/sicejl1962.42.631 　
Hyperparameter Estimation of Multi-valued Image Restoration by means of Q-state Ising Model

Tanaka Kazuyuki, Inoue Junichi

Meeting abstracts of the Physical Society of Japan　58　(1)　257-257　2003/03/06
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Dynamics of hyper-parameter estimation via EM algorithrn in image restoration using the Q-lsing model

Inoue Junichi, Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　58　(1)　257-257　2003/03/06
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Image compression and decompression by using a mean field approximation for an elastic net

Ito Y., Tanaka K., Horiguchi T.

Meeting abstracts of the Physical Society of Japan　58　(1)　256-256　2003/03/06
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Belief Propagation Algorithm for Bayesian Network by means of Cluster Variation Method and Linear Response Theory

Tanaka Kazuyuki

Proceedings of the IEICE General Conference　2003　(1)　9-9　2003/03/03
Publisher: The Institute of Electronics, Information and Communication Engineers
ミニ特集/ベイズ統計･統計力学と情報処理

田中和之

計測と制御　2003
Statistical-mechanical approach to image processing

K Tanaka

JOURNAL OF PHYSICS A-MATHEMATICAL AND GENERAL　35　(37)　R81-R150　2002/09

DOI： 10.1088/0305-4470/35/37/201 　

ISSN： 0305-4470
Belief Propagation for Probabilistic Information Processing and Cluster Variation Method

TANAKA Kazuyuki

IEICE technical report. Neurocomputing　101　(737)　199-206　2002/03/13
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

General formula for berief propagation algorithms are given by applying a cluster variation mehtod to massive probabilistic models with any network structure. We give some numerical experiments for the application of the belief propagation algorithm to binary image processing and its hyperparameter determination constructed by means of Bayesian statistics.
24pXY-2 Berief Propagation for Probabilistic Information Processing and Cluster Variation Method

Tanaka Kazuyuki

Meeting abstracts of the Physical Society of Japan　57　(1)　242-242　2002/03/01
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Automatic parameter estimation in probabilistic image processing system based on Gibbs microcanonical distribution

Tanaka Kazuyuki

Proceedings of the Society Conference of IEICE　2001　8-8　2001/08/29
Publisher: The Institute of Electronics, Information and Communication Engineers
Bayesian statistical approach to hyper-parameter estimation by using solvable Markov random field model

Inoue J., Tanaka K.

Meeting abstracts of the Physical Society of Japan　56　(1)　271-271　2001/03/09
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Hyperparameter Estimation of Solvable Markov Random Field Model in Image Restoration

Tanaka Kazuyuki, Inoue Jun-ichi

Proceedings of the IEICE General Conference　2001　(1)　18-18　2001/03/07
Publisher: The Institute of Electronics, Information and Communication Engineers
Staistical-Mechanical Interative Algorithm by Means of Cluster Variation Method in Compound Gauss-Markov Random Field Model

16　(2)　322-322　2001/03/01

ISSN： 0912-8085
Bayesian-statistical Approach in Image Restoration by means of Solvable Markov Random Field Model

TANAKA Kazuyuki, INOUE Jun-ichi

Technical report of IEICE. PRMU　100　(507)　41-48　2000/12/07
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

We give the exact expression of the marginal likelihood in a solvable Markov random field model which have a form of Gibbs distribution. The solvable Markov random field model is equivalent to the Wiener filter of the simultaneous auto-regressive model. The estimates of hyperparameters are obtained by maximizing the marginal likelihood. We study the statistical property of the solvable Markov random field model, exactly.
Dynamical properties of algorithms for image restoration by means of Bayesian statistics

INOUE Jun-ichi, TANAKA Kazuyuki

Technical report of IEICE. PRMU　100　(507)　33-40　2000/12/07
Publisher: The Institute of Electronics, Information and Communication Engineers
ISSN： 0913-5685

More details Close

Recently, properties of image restoration were investigated in the context of Bayesian approach. However, these results are restricted to the static properties of the algorithms and no studies have ever tried to investigate these dynamical properties explicitly. In this report, we introduce an exactly solvable model for image restoration and derive the differential equation with respect to macroscopic quantities(Hamming distance between original and restored images, etc.)analytically. From these dynamical equations, we obtain useful information for image restoration, for example, basin of attraction, speed of convergence, etc. Our approach also enable one to investigate the hyper-parameter estimation by means of maximization of marginal likelihood using steepest descent from dynamical point of view.
Approximate Optimization Algorithms in Markov Random Field Model Based on Statistical-Mecanical Techniques (<Special Issue>International Workshop on Soft Computing in Industry '99(IWSCI'99))

TANAKA Kazuyuki, MAEDA Junji

Memoirs of the Muroran Institute of Technology　50　9-15　2000/11/30
Publisher: Muroran Institute of Technology
ISSN： 1344-2708
Statistical mechanical analysis of performance for hyperparameter estimation by means of maximization of marginal likelihood

Inoue J., Tanaka K.

Meeting abstracts of the Physical Society of Japan　55　(2)　255-255　2000/09/10
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Image Restoration and Compound Gauss-Markov Random Field Model with Quantum-Mechanical Edge State

TANAKA Kazuyuki

Meeting abstracts of the Physical Society of Japan　55　(2)　255-255　2000/09/10
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Statistical analysis of performance for hyperparameter estimation by means of maximization of marginal likelihood in Bayesian statistics

Proceedings of the Society Conference of IEICE　2000　20-20　2000/09/07
Publisher: The Institute of Electronics, Information and Communication Engineers
Image Restoration and Markov Random Field Model with Quantum-Mechanical Superposed Edge State

TANAKA Kazuyuki

Proceedings of the Society Conference of IEICE　2000　19-19　2000/09/07
Publisher: The Institute of Electronics, Information and Communication Engineers
Probabilistic Model and Image Processing

TANAKA Kazuyuki

Meeting abstracts of the Physical Society of Japan　55　(1)　243-243　2000/03/10
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Probabilistic Algorithm in Compound Gauss-Markov Random Field Model by Cluster Variation Method

TANAKA Kazuyuki

Proceedings of the IEICE General Conference　2000　(1)　25-25　2000/03/07
Publisher: The Institute of Electronics, Information and Communication Engineers
Probabilistic Computing and Image Processing

TANAKA Kazuyuki

12　141-146　2000/01/21
Overview of the Tsukuba Advanced Research Alliance (TARA) research initiative on quality of life of older Japanese adults

K. Tanaka, M. Okada

Journal of Aging and Physical Activity　8　(2)　109-112　2000
Publisher: Human Kinetics Publishers Inc.
DOI： 10.1123/japa.8.2.109 　

ISSN： 1063-8652
Image Restiration and Statistical Physics

Kazuyuki Tanaka

Mathematical Sciences　37　(12)　20-27　1999/12/01
Publisher: SAIENSU-SHA Co.,Ltd
27pU-6 Image Restoration by Compound Gauss-Markov Random Field Model and Cluster Variation Method

TANAKA Kazuyuki

Meeting abstracts of the Physical Society of Japan　54　(2)　285-285　1999/09/03
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Statistical-mechanical Study of Maximum Marginal Likelihood Estimation in Markov Random Field Model

TANAKA Kazuyuki

Proceedings of the Society Conference of IEICE　1999　10-10　1999/08/16
Publisher: The Institute of Electronics, Information and Communication Engineers
30p-PSA-9 Effect of energy functions for restoration of multi-value image II

Miya M, Tanaka K, Horiguchi T

Meeting abstracts of the Physical Society of Japan　54　(1)　679-679　1999/03/15
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
31p-XC-3 Statistical-mechanical Methods in Parameter Estimation of Markov Random Field Model

TANAKA Kazuyuki

Meeting abstracts of the Physical Society of Japan　54　(1)　697-697　1999/03/15
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
Maximum Posterior Marginal Estimation in Image Restoration

TANAKA Kazuyuki

Proceedings of the IEICE General Conference　1999　(1)　48-48　1999/03/08
Publisher: The Institute of Electronics, Information and Communication Engineers
Statistical-Mechanical Method in Image Restoration

Kazuyuki Tanaka

BUTSURI　54　(1)　25-33　1999/01/05
Publisher: 社団法人日本物理学会
DOI： 10.11316/butsuri1946.54.25 　

ISSN： 0029-0181
Statistical-mechanical Study of Maximum Likelihood Hyperparameter Estimation in Markov Random Field Model

TANAKA Kazuyuki

Proceedings of the Society Conference of IEICE　1998　72-72　1998/09/07
Publisher: The Institute of Electronics, Information and Communication Engineers
Edge Ditection for Gray-Level Image based on Ising Model

TANAKA Kazuyuki, MORITA Tohru

Meeting abstracts of the Physical Society of Japan　53　(2)　754-754　1998/09/05
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
28p-YH-7 Cluster Zero-temperature Process in the Recovery of an Image

Morita T., Tanaka K.

Meeting abstracts of the Physical Society of Japan　52　(1)　680-680　1997/03/17
Publisher: The Physical Society of Japan (JPS)
ISSN： 1342-8349
A Formulation of Mean-Field Annealing Argorithm in Image Restoration Problem based on Marcov Random Field Model

TANAKA Kazuyuki, MORITA Tohru

Memoirs of the Muroran Institute of Technology. Science and engineering　46　9-25　1996/11
Publisher: Muroran Institute of Technology
ISSN： 0580-2415
Statistical-Mechanical Iterative Argorithm for Recovery of Image with Grades

TANAKA Kazuyuki, MORITA Tohru

Proceedings of the Society Conference of IEICE　1996　234-234　1996/09/18
Publisher: The Institute of Electronics, Information and Communication Engineers
Annealing proceedure and image restoration based on Potts model

Tanaka K., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1996　(3)　722-722　1996/09/13
Publisher: The Physical Society of Japan (JPS)
1a-L-12 Cluster variation method and image restoration based on Potts model

Tanaka K., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Annual meeting　51　(3)　593-593　1996/03/15
Publisher: The Physical Society of Japan (JPS)
Cluster Variation Method and Markov Random Field Model

Tanaka Kazuyuki, Morita Tohru

Proceedings of the IEICE General Conference　1996　(2)　164-164　1996/03/11
Publisher: The Institute of Electronics, Information and Communication Engineers
Determination of Parameters in an Image Restoration by Statistical-mechanical Means

Morita T., Tanaka K.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1995　(3)　697-697　1995/09/12
Publisher: The Physical Society of Japan (JPS)
Cluster variation method and image restoration based on Ising models II

Tanaka K., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1995　(3)　698-698　1995/09/12
Publisher: The Physical Society of Japan (JPS)
Quantum lattice gases equivalent to the Heisenberg model and the t-J model

Morita T., Tanaka K.

Abstracts of the meeting of the Physical Society of Japan. Annual meeting　50　(3)　651-651　1995/03/16
Publisher: The Physical Society of Japan (JPS)
2a-D-2 One-dimensional VBS model with spin greater than 1/2

Morita T, Tanaka K

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1994　(3)　421-421　1994/08/16
Publisher: The Physical Society of Japan (JPS)
5a-D-6 Cluster variation method for Hubbard model and t-J model II

Tanaka K, Morita T

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1994　(3)　526-526　1994/08/16
Publisher: The Physical Society of Japan (JPS)
29p-PSA-35 Spin dependence of non-universal phase transition

Honda Y, Tanaka K, Horiguchi T

Abstracts of the meeting of the Physical Society of Japan. Annual meeting　49　(3)　555-555　1994/03/16
Publisher: The Physical Society of Japan (JPS)
29p-PSA-7 Generalized Heisenberg model and t-J model of spin more than 1/2

Morita T, Tanaka K

Abstracts of the meeting of the Physical Society of Japan. Annual meeting　49　(3)　543-543　1994/03/16
Publisher: The Physical Society of Japan (JPS)
29p-PSA-13 Cluster variation method for Hubbard model and t-J model

Tanaka Kazu, Ebisawa H, Morita T

Abstracts of the meeting of the Physical Society of Japan. Annual meeting　49　(3)　545-545　1994/03/16
Publisher: The Physical Society of Japan (JPS)
12a-H-6 Monte Carlo renormalization-group study on the two dimensional spin-1 Ising model with frustration

Honda Y., Tanaka K., Saito S.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1993　(3)　523-523　1993/09/20
Publisher: The Physical Society of Japan (JPS)
12a-H-1 Numerical study of 2-d ANNNI model in the cluster variation method

Murai Y., Tanaka K., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1993　(3)　520-520　1993/09/20
Publisher: The Physical Society of Japan (JPS)
12a-H-7 S=1 Ising model with next-nearest-neighbor interaction and cluster variation method

Tanaka K., Honda Y., Horiguchi T., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1993　(3)　523-523　1993/09/20
Publisher: The Physical Society of Japan (JPS)
15a-H-12 Exactly solved one-dimensional quantum spin system with spin more than one

Tanaka K., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1993　(3)　627-627　1993/09/20
Publisher: The Physical Society of Japan (JPS)
15a-H-13 Heisenberg model of spin more than 1/2 equivalent to the one of spin 1/2

Morita T., Tanaka K.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1993　(3)　628-628　1993/09/20
Publisher: The Physical Society of Japan (JPS)
14p-H-9 Hubbard model and Cluster variation method

Tanaka K., Ebisawa H., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Sectional meeting　1993　(3)　612-612　1993/09/20
Publisher: The Physical Society of Japan (JPS)
30a-E-6 Three-dimensional t-J model and cluster variation method

Tanaka K., Takane Y., Ebisawa H., Morita T.

Abstracts of the meeting of the Physical Society of Japan. Annual meeting　48　(3)　483-483　1993/03/16
Publisher: The Physical Society of Japan (JPS)
26a-ZB-1 Diagram Method for 2-D Ising Model-to Free-Fermion Vertex Model

Morita T, Tanaka K

1992　(3)　459-459　1992/09/14
Publisher: The Physical Society of Japan (JPS)
26a-ZB-2 Asymptotic Behavior of Spin-Pair Correlation Function of Ising Model on Checkerboard Lattice

Tanaka Kazuyuki, Morita Tohru

1992　(3)　460-460　1992/09/14
Publisher: The Physical Society of Japan (JPS)
26a-ZB-3 Integration of Lattice Sites for Two-Sublattice Free-Fermion Eight-Vertex Model

Morita T, Tanaka K

1992　(3)　460-460　1992/09/14
Publisher: The Physical Society of Japan (JPS)
28a-ZM-6 2D Ising Model with Competing Interactions and Cluster Variation Method II

Tanaka K, Horiguchi T, Morita T

47　(3)　443-443　1992/03/12
Publisher: The Physical Society of Japan (JPS)
28p-BPS-37 2D Ising Model with Competing Interactions and Cluster Variation Method

Tanaka K., Horiguchi T., Morita T.

46　(3)　466-466　1991/09/12
Publisher: The Physical Society of Japan (JPS)
25a-J-6 Low-temperature behavior of anisotropic antiferromagnetic Ising model on triangular lattice

Tanaka K., Horiguchi T., Morita T.

1991　(3)　452-452　1991/03/11
Publisher: The Physical Society of Japan (JPS)
2次元イジング模型のスピン相関関数とテープリッツ行列式(基研研究会「新しい統計物理学の基礎 : 多様性の中の類似性」,研究会報告)

田中和之, 守田徹

物性研究　57　(1)　68-73　1991
Publisher: 物性研究刊行会
ISSN： 0525-2997
3p-Q-12 Critical exponent η of Ising model in cluster variation method and CAM

Tanaka K., Horiguchi T., Morita T.

1990　(3)　461-461　1990/10/02
Publisher: The Physical Society of Japan (JPS)
4p-PS-16 Asymptotic Behavior of the Spin Pair Correlation Functions for Two-Dimensional Ising Models

Tanaka K., Morita T., Hiroike K.

1989　(3)　416-416　1989/09/12
Publisher: The Physical Society of Japan (JPS)
6p-D5-2 レンガ格子上のANNNI模型におけるスピン相関関数

田中和之, 守田徹, 廣池和夫

秋の分科会講演予稿集　1988　(3)　428-428　1988/09/16
Publisher: 一般社団法人日本物理学会
29p-W-4 非加算的直径を持つ剛体球の1次元2成分気体 I

守田徹, 田中和之, 廣池和夫

秋の分科会講演予稿集　1987　(3)　385-385　1987/09/16
Publisher: 一般社団法人日本物理学会
28p-QB-9 1次元格子気体の直接相関関数 II

田中和之, 守田徹, 廣池和夫

秋の分科会講演予稿集　1986　(3)　345-345　1986/09/12
Publisher: 一般社団法人日本物理学会
1a-D-5 1次元格子気体の直接相関関数

田中和之, 廣池和夫, 守田徹

秋の分科会講演予稿集　1985　(3)　264-264　1985/09/13
Publisher: 一般社団法人日本物理学会

Show all ︎Show first 5

Books and Other Publications 15

Sublinear Computation Paradigm ---Algorithmic Revolution in the Big Data Era---

Naoki Katoh, Yuya Higashikawa, Hiro Ito, Atsuki Nagao, Tetsuo Shibuya, Adnan Sljoka, Kazuyuki Tanaka, Yushi Uno

Springer　2021/10
Statistical Modeling on Image Processing ---Approaches from Probabilistic Model and Sparse Modeling---

Shun Kataoka, Masayuki Ohzeki, Muneko Yasuda, Kazuyuki Tanaka

Kyoritsu Shuppan Co., Ltd.　2018/11

ISBN: 9784320111233
Encyclopedia of Artificial Intelligence

Shun Kataoka, Muneko Yasuda, Kazuyuki Tanaka

Kyoritsu Co. Ltd.　2017/07

ISBN: 9784320124202
Probabilistic Graphical Model

Jyo Suzuki, Naomi Ueno, Manabu Kuroki, Shohei Shimizu, Shin-ichi Minato, Masakazu Ishihata, Yoshiyuki Kabashima, Kazuyuki Tanaka, Yoichi Motomura, Yoshinori Tamada

Kyoritsu Shuppan Co., Ltd.　2016/07
Computer Vision and Image Media 3 --Bundle Adjustment, ICP, Bag-of-Features, Nearest Neighbor Search, Stochastic Image Processing--

Takayuki Okatani, Ken Masuoka, Koichi Kise, Keiji Yanai, Toshikazu Wada, Muneki Yasuda, Shun Kataoka, Kazuyuki Tanaka

Advanced Communication Media CO., Ltd.　2010/12

ISBN: 9784915851407
International Workshop on Statistical-Mechinical Informatics 2010 (Journal of Physics, Conference Series, Vol.233)

Jun-ichi Inoue, Yoshiyuki Kabashima, Kazuyuki Tanaka, Toshiyuki Tanaka

IOP Publishing　2010/07/19
映像情報メディア工学大事典(映像情報メディア学会編)/基礎編, 第4部門. 画像処理, 第12章. 確率的画像処理

Kazuyuki Tanaka, Muneki Yasuda

オーム社　2010/06/15
Mathematics in Statistical Inference of Bayesian Networks

Kazuyuki Tanaka

Corona Publishing Co., LTD.　2009/10/28

ISBN: 9784339024418
International Workshop on Statistical Mechanical Informatics 2008 (Journal of Physics: Conference Series, Vol.143)

Masahito Hayashi, Jun-ichi Inoue, Yoshiyuki Kabashima, Kazuyuki Tanaka

IOP Publishing　2009/01/07
Applied Mathematics for Electronics and Information Sciences

Kazuyuki Tanaka, Masahiko Hayashi, Hiromichi Ebisawa

Asakura Publishing Co.　2007/04/15
Probabilistic Information Processing and Statistical Mechanics

Kazuyuki Tanaka

Saiensu Publishing Co.　2006/09/25
Introduction of Probabilistic Model and Image Processing

Kazuyuki Tanaka

森北出版　2006/09/25
Encyclopedia if Artificial Intelligence/14-12. Markov Random Field

Kazuyuki Tanaka, Edited by, Japanese Artificial Intelligence Society

Kyoritsu Shuppan CO.　2005/12/10
Statistical Physics of Disordered Systems and Its Applications

Koji Hukushima, Kazuyuki Tanaka, Hidetoshi Nishimori

Progress of Theoretical Physics Supplement (Published by the Yukawa Institute of Theoretical Physics and The Physical Society of Japan)　2005/04/30
Theory and Applications of the Cluster Variation and Path Probability Methods/Application of the Cluster Variation Method to Image Restoration Problem

Kazuyuki Tanaka, Tohru Morita

Plenum Press　1996

Show all Show first 5

Presentations 30

Adaptive Quantum Monte Carlo method for a class of non-stoquastic Hamiltonian by using D-Wave machine

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

The Third D-Wave Qubits North America Users Conference (25-28 September, 2018, Knoxville, TN 37902, USA)
Statistical-mechanical analysis of restricted Boltzmann machine with transverse field

Chako Takahashi, Masayuki Ohzeki, Kazuyuki Tanaka

Adiabatic Quantum Computing 2019 (AQC-19)　2019/07/26
Mean-field analysis of quantum error-correcting codes with non-stoquastic Hamiltonian

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

Adiabatic Quantum Computing 2019 (AQC-19)　2019/07/26
Acceleration of adaptive quantum Monte Carlo Sampling for a class of non-stoquastic Hamiltonian by using D-Wave 2000Q

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

Adiabatic Quantum Computing Conference 2018 (25-28 June, 2018, NASA Conference Center, Building 3 Moffett Field, CA 94035, USA)
Deep neural network detects quantum phase transition in d-wave 2000Q

Shunta Arai, Masayuki Ohzeki, Kazuyuki Tanaka

Adiabatic Quantum Computing Conference 2018 (25-28 June, 2018, NASA Conference Cen ter, Building 3 Moffett Field, CA 94035, USA)
Plefka Expansion of Replica Extended Ising Model and Its Application to Inverse Ising Problem International-presentation

Muneki Yasuda, Yoshiyuki Kabashima, Kazuyuki Tanaka

24 IUPAP International Conference on Statistical Physics　2010/07/13
The Approximate Linear Response Theory and Its Mathmatical Structures

Muneki Yasuda, Kazuyuki Tanaka

23 IUPAP International Conference on Statistical Physics　2007/07/09
Advanced Mean Field Methods in Quantum Probabilistic Inference

Kazuyuki Tanaka

23 IUPAP International Conference on Statistical Physics　2007/07/09
ガウシアングラフィカルモデルを用いた画像補修アルゴリズム

安田宗樹, 大久保潤, 田中和之

第5回情報科学技術フォーラム　2006/09/05

More details Close

安田宗樹が登壇した．
混合のエントロピーを利用したコミュニティ検出アルゴリズム

大久保潤, 田中和之

第5回情報科学技術フォーラム　2006/09/05

More details Close

大久保潤が登壇した．
Statistical Learning Procedure in Loopy Belief Propagation for Probabilistic Image Processing International-presentation

International Conference on Computational Intelligence for Modelling, Control and Automation　2005/11/28
Linear Response Formula in Generalized Belief Propagation for Probabilistic Inference International-presentation

Kazuyuki Tanaka

International Conference on Computational Intelligence for Modelling, Control and Automation　2005/11/28
Digital Images Inpainting based on Markov Random Field International-presentation

Muneki Yasuda, Jun Ohkubo, Kazuyuki Tanaka

International Conference on Computational Intelligence for Modelling, Control and Automation　2005/11/28
Preferential Urn Model on a Random Network International-presentation

Jun Ohkubo, Kazuyuki Tanaka

International Conference on Computational Intelligence for Modelling, Control and Automation　2005/11/28
Graphical model and generalized belief propagation in probabilistic image processing International-presentation

Kazuyuki Tanaka

Statistical Physics of Disordered Systems and Its Applications　2005/09/05
Probabilistic image processing based on the Q-Ising model by means of the mean-field method and loopy belief propagation International-presentation

Kazuyuki Tanaka, D. M. Titteringto

17th International Conference on Pattern Recognition　2004/08/23

More details Close

(Proceedings) Kazuyuki Tanaka and D. M. Titterington: Probabilistic image processing based on the Q-Ising model by means of the mean-field method and loopy belief propagation, Proceedings of 17th International Conference on Pattern Recognition, Vol.2, pp.40-43, 2004, IEEE Computer Society Press.
First-order phase transition and Bayesian image processing by loopy belief propagation International-presentation

Kazuyuki Tanaka, D. M. Titteringto

Satellite Meeting of STATPHYS 22 "Statistical Physics of Disordered Systems and Its Applications"　2004/07/12

More details Close

(Proceedings) Kazuyuki Tanaka and D. M. Titterington: First-order phase transition and Bayesian image processing by loopy belief propagation, Progress of Theoretical Physics Supplement, No.157, pp.288-291, 2005.
Dynamics and its stability of Boltzmann-machine learning algorithm for gray scale image restoration International-presentation

Jun-ichi Inoue, Kazuyuki Tanaka

3rd International Symposium on Slow Dynamics in Complex Systems　2003/11/03

More details Close

(Proceedings) Jun-ichi Inoue and Kazuyuki Tanaka: Dynamics and its stability of Boltzmann-machine learning algorithm for gray scale image restoration, Proceedings of 3rd International Symposium on Slow Dynamics in Complex Systems (edited by M. Tokuyama and I. Oppenheim), pp.731-734 (American Institute of Physics Press).
Loopy belief propagation and probabilistic image processing International-presentation

Kazuyuki Tanaka, Jun-ichi Inoue, D. M. Titteringto

2003 IEEE Workshop on Neural Networks for Signal Processing　2003/07/17

More details Close

(Proceedings) Neural Networks for Signal Processing XIII ---Proceedings of the 2003 IEEE Signal Processing Society Workshop, pp.329-338, 2003, IEEE Computer Society Press.
Statistical-mechanical approach to probablistic image processing ---Loopy belief propagation and advanced mean-field method--- International-presentation

Kazuyuki Tanaka, Noriko Yoshiike

2003 Joint Workshop of Hayashibara Foundation and Grant-in Aid for Scientific Research on Priority Areas: Statistical Mechanical Approach to Probabilistic Information Processing, "Physics and Information"　2003/07/11
Cluster Variation Method for Correlation Function of Probabilistic Model with Loopy Graphical Structure International-presentation

Kazuyuki Tanaka

2002 Neural Information Processing Systems (NIPS*2002) Workshop on "Propagation Algorithms on Graphs with Cycles: Theory and Applications"　2002/12/13
Dynamical properties of image restoration and hyper-parameter estimation International-presentation

Jun-ichi Inoue, Kazuyuki Tanaka

2001 IEEE Workshop on Newral Networks for Signal Processing　2001/09/10

More details Close

(Proceedings) Jun-ichi Inoue and Kazuyuki Tanaka: Dynamical properties of image restoration and hyper-parameter estimation, Neural Networks for Signal Processing XI ---Proceedings of the 2001 IEEE Signal Processing Society Workshop, pp. 383-392, 2001.
Dynamics of image restoration and hyper-parameter estimation International-presentation

Jun-ichi Inoue, Kazuyuki Tanaka

21st IUPAP International Conference on Statistical Physics　2001/07/15
Image Restoration based on Constrained Optimization and Classical Spin Systems International-presentation

Kazuyuki Tanaka, Tohru Morita

20st IUPAP International Conference on Statistical Physics　1998/07/15
Statistical-Mechanical Algorithm in MRF Model based on Variational Principle International-presentation

Kazuyuki Tanaka, Mikako Ichioka, Tohru Morita

13th International Conference on Pattern Recognition　1996/08/25

More details Close

(プロシーディング) Proceedings of the 13th International Conference on Pattern Recognition, Track B: Pattern Recognition, IEEE Computer Society Press, pp.381-388, 1996.
Correlated Random Field Ising Model and Image Restoration Problem International-presentation

Kazuyuki Tanaka, Tohru Morita

International Symposium on Coferent Approaches to Fluctuations　1995/07/17

More details Close

(プロシーディング) Proceedings of International Symposium on Coferent Approaches to Fluctuations, World Scientific Press, pp.150-153, 1995.
Application of Cluster Variation Method to Image Restoration Problem International-presentation

Kazuyuki Tanaka, Tohru Morita

International Workshop on the Theory and Applications of the Cluster Variation and Path Probability Methods　1995/06/18

More details Close

(プロシーディング) Theory and Applications of the Cluster Variation and Path Probability Methods Plenum Press, pp.353-373, 1996.
Application of the Culster Variation Method to Three-Dimensional t-J model II International-presentation

Kazuyuki Tanaka, Tohru Morita

1994 Workshop on Statistical Physics　1994/09/29

More details Close

(プロシーディング) Journal of the Korean Physical Society (Supplementary Issue), Vol.28, pp.S373-S377, 1994
Generalized Heisenberg model of spin S Greater than 1/2 equivalent to those of spin 1/2 or 1 International-presentation

Tohru Morita, Kazuyuki Tanaka

1994 Workshop on Statistical Physics　1994/09/29

More details Close

(プロシーディング) Journal of the Korean Physical Society (Supplementary Issue) Vol.28, pp.S378-S383, 1994.
Two-dimensional layered Ising model International-presentation

Kazuyuki Tanaka, Tohru Morita

18th IUPAP International Conference on Statistical Physics　1992/08/02

Show all Show first 5

Industrial Property Rights 1

能力推定システムおよび方法ならびにプログラムおよび記憶媒体

菅谷史昭, 田中和之

5029090

Property Type: Patent

Research Projects 19

Realizations and practical applications of quantum statistical machine learning theory with navigation functions

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research

Category: Grant-in-Aid for Scientific Research (B)

Institution: Tohoku University

2022/04/01 - 2025/03/31
量子統計的機械学習システム設計理論の創出とその実践 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金基盤研究(B)

2018/04 - 2021/03
Extraction of latent structure by sparse modeling

Okada Masato

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research

Category: Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area)

Institution: The University of Tokyo

2013/06/28 - 2018/03/31

More details Close

The sparse modeling team (B01-2) sets three tasks. Task 1 is applications of Bayesian spectral decomposition method to actual data. We developed a noise variance estimation method and a fast calculation method using L1 regularization and verified its effectiveness with actual data. In task 2, we developed a basis estimation and selection method using Sp-DMD for time series data, and applied it to actual data and verified its effectiveness. In task 3, a method of evaluating the appropriateness of the basis combination using an exhaustive search was developed, and this method was applied to actual data and the effectiveness was verified. Through research on these three tasks, we developed a universal method to extract latent structures using SpM and verified its effectiveness by actual data.
非加算的体積とエントロピーによるコミュニティ抽出の確率的計算モデル設計理論 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金基盤研究(B)

2013/04 - 2017/03
量子力学的ゆらぎを伴うベイジアンネットシステムの設計理論の構築 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金基盤研究(C)

2005/04 - 2017/03
Scene Text Recognition without Text Region Estimation

Waizumi Yuji, TANAKA KAZUYUKI

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research

Category: Grant-in-Aid for Challenging Exploratory Research

Institution: Tohoku University

2014/04/01 - 2016/03/31

More details Close

The purpose of this work is building an algorithm for scene text recognition without text region estimation. Almost previous works for scene text recognition consist of two recognition stages, that is text location detection and character recognition. Undetected texts at the text location detection stage are never recognized in character recognition stage. This is a serious drawback of the two stage recognition algorithm. In this work, we propose a new scene text recognition algorithm without text location detection stage to overcome the drawback. The proposed algorithm considers an scene image as a set of closed curves and all closed curves are a part of a letter. By applying score function to all closed curves in a scene image, the proposed method can reduce un-recognition for the undetected regions.
マルコフ確率場によるベイジアンアドホックネットワークシステム設計理論の創出 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金

2012/04 - 2014/03
量子力学的ラベリングにより拡張されたデータマイニング理論の創出 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金基盤研究(B)

2010/04 - 2013/03
Web Application Identification Based on Statistical Estimation of Unobserved Information

WAIZUMI Yuji, TANAKA Kazuyuki

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research

Category: Grant-in-Aid for Scientific Research (C)

Institution: Tohoku University

2011 - 2013

More details Close

We developed an identification method for Web application provided from a Web site using network traffic only. Identifying method cannot use port number or IP address because all traffic from a Web site are blocked if packets which have the IP address of the Web site. So we proposed a new identification method which uses only packet sizes and the number of TCP connection. The identification accuracy achieves 90%.
大規模確率場により高度化された確率推論システムの設計 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金特定領域研究

2006/04 - 2010/03
Management Research on "Deepening and Expansion of Statistical Mechanical Informatics"

KABASHIMA Yoshiyuki, OKADA Masato, TANAKA Kazuyuki, TANAKA Toshiyuki, ISHII Shin, INOUE Junichi

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research

Category: Grant-in-Aid for Scientific Research on Priority Areas

Institution: Tokyo Institute of Technology

2006 - 2009

More details Close

This research project played a role of managing the priority area "Deepening and Expansion of Statistical Mechanical Informatics". More precisely, we formulated the grand design of research directions, coordinated individual research projects in the priority area, organized public symposia and international conferences, promoted research achievements and made evaluation and advice to the individual projects. Major records in the research duration include organization of four public symposia and six international conferences, and publication of four post conference proceedings. The significance of the activity is reflected in publication of more than 280 original papers from this area.
確率的情報処理への統計力学的アプローチに関する総括研究 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金特定領域研究

2002/04 - 2007/03
マルコフ確率場モデルに基づく適応型画像処理フィルターの設計理論 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金特定領域研究

2002/04 - 2006/03
量子力学的確率モデルによる知識情報処理におけるトンネル効果の応用 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金基盤研究(C)

2001/04 - 2003/03
知識情報処理への統計力学的アプローチ Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金基盤研究(C) 企画調査

2000/04 - 2001/03
画像処理における統計力学的概念をもとにしたソフトコンピューティング理論の構築 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金奨励研究(A)

1998/04 - 2000/03
STUDY ON MULTI-MEDIA DATABASE USING FUZZY ASSOCIATIVE MEMORY

SUZUKI Yukinori, SATO Saga, MAEDA Junji

Offer Organization: Japan Society for the Promotion of Science

System: Grants-in-Aid for Scientific Research

Category: Grant-in-Aid for Scientific Research (B)

Institution: Muroran Institute of Technology

1998 - 1999

More details Close

We studied an associative memory system (AMS) for a multi-media database. The AMS is designed using fuzzy numbers, which consists of a memory matrix and a defuzzification procedures. The memory matrix associates fuzzy response vectors which are transformed into non-fuzzy response vectors by the 3defuzzification procedure. The performance of the AMS was evaluated by computational experiments. The results of the experiments indicated that the memory capacity of the AMS was larger than that of the standard linear associative memories. In addition, the noise tolerance of the AMS was practically the same as that of the linear associative memory using the generalized inverse amtix method. As a result, there is a potential to use the AMS for database.
クラスター変分法とマルコフ確率場モデルをもとにした確率的学習・探索機構の構築 Competitive

田中和之

Offer Organization: 文部科学省

System: 科学研究費補助金奨励研究(A)

1996/04 - 1997/03
超伝導体を含む微細構造における電子間相互作用と位相干渉の基礎理論

海老澤丕道, 田中和之

Offer Organization: 日本学術振興会

System: 科学研究費助成事業

Category: 重点領域研究

Institution: 東北大学

1994 - 1994

More details Close

超伝導・常伝導の複合系でいずれもがメゾスコピックな領域にあるときの両者の位相のかかわりに着目し、 (1)境界での準粒子の透過・反射・アンドレエフ反射等を考慮して、そのような体系のコンダクタンスのふるまいを評価する、 (2)超伝導転移の理論的取り扱い方を考察する、 (3)電子間相互作用の効果を考える、の3点が本研究の目標であった。本年度の実績は、そのうち第1点についての結果が得られたことである。超伝導体と接触している常伝導体におけるコンダクタンスとその揺らぎには、そのサイズに依存して特徴ある影響が現れる。2次元モデル系の電子透過率を、波動関数を逐次に計算して求めるやり方で数値的に評価し、コンダクタンスの特徴について考察した。常伝導(N)と超伝導(S)部分のSNSNモザイク構造に対する結果は、超伝導部分の長さが超伝導コヒーレンス長の程度で示すクロスオーバーの特徴を示し、アンドレーフ反射が完全であるか、また超伝導電流への転換が完全であるかが重要であることを示すものであり、これは基本的で重要な知見である。

Show all Show first 5

Social Activities 4

Computational Intelligence Seminar in Japan (IEEE Computational Intelligence Society Japan Chapter 主催企画)

2005/11/08 -

More details Close

確率推論を軸として，確率伝搬法，進化計算，確率的情報処理，理論と応用という観点からの大学院学生，大学・企業の若手研究者向けのわかりやすい解説と最新の研究紹介を行う講演会を企画し，実行委員を務めた．講演者（敬称略）： Yair Weiss (The Hebrew University of Jerusalem, Israel) 田中和之 (東北大学大学院情報科学研究科) 田中利幸 (京都大学大学院情報学研究科)
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「確率的アルゴリズムによる情報処理」

2004/11 -

More details Close

確率的情報処理手法の一つであるモンテカルロ法についての最新技術を基礎的なメカニズムから具体的な情報処理への応用事例まで含めた初心者向けのわかりやすい紹介を行う講演会を企画し，世話人代表を務めた．開催場所：ぱるるプラザ京都 6 階会議室 C 講演者（敬称略）：福島孝治 (東京大学大学院総合文化研究科) 喜多一 (京都大学学術情報メディアセンター) 渡辺治 (東京工業大学大学院情報理工学研究科)
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「確率的情報処理としての移動体通信技術」

2003/12 -

More details Close

役割：世話人代表．開催趣旨：CDMA およびターボ符号 (より一般には近似的反復復号アルゴリズム) を例にとり，分野複合的な視点から最新の研究成果をわかりやすく解説する講演会を企画し，世話人代表をつとめた．講演者（敬称略）：池田思朗 (九州工業大学大学院生命体工学研究科) 田中利幸 (東京都立大学大学院工学研究科) 岡田真人 (理化学研究所脳科学総合研究センター)
電子情報通信学会基礎・境界ソサイエティ主催企画若手研究者・学生向けに最新技術をわかりやすく紹介する講演会「ベイズ統計と統計力学を用いた確率的情報処理技術」

2002/03/22 -

More details Close

ベイズ統計にもとづく画像処理技術，情報通信技術，知識情報処理技術の確率モデル化，統計力学的手法を用いたアルゴリズム構成手法，確率的情報処理システムの解析的性能評価についての大学院学生，大学・企業の若手研究者を対象とする初心者向けのわかりやすい紹介を行う講演会を企画し，世話人を務めた．講演者：西森秀稔（東京工業大学大学院理工学研究科）樺島祥介 (東京工業大学大学院総合理工学研究科) 田中和之 (東北大学大学院情報科学研究科)

Other 1

ビッグデータ時代に向けた革新的アルゴリズム基盤

More details Close

本研究の目的は，ビッグデータ時代に向けた新しい計算パラダイム「劣線形時間パラダイム」を提唱し，そのためのアルゴリズムとデータ構造，モデリング技法を開発し，ビッグデータ時代に向けた革新的アルゴリズム基盤を構築することである．