Fuzzy Multiple Discrimminant Analysis (FMDA) 5) (SOM) 6) SOM 3 6) SOM SOM SOM SOM SOM SOM 7) 8) SOM SOM SOM GPU 2. n k f(x) m g(x) (1) 12) { min(max)

Similar documents
GPGPU

23 Fig. 2: hwmodulev2 3. Reconfigurable HPC 3.1 hw/sw hw/sw hw/sw FPGA PC FPGA PC FPGA HPC FPGA FPGA hw/sw hw/sw hw- Module FPGA hwmodule hw/sw FPGA h

1 Table 1: Identification by color of voxel Voxel Mode of expression Nothing Other 1 Orange 2 Blue 3 Yellow 4 SSL Humanoid SSL-Vision 3 3 [, 21] 8 325

A Study on Throw Simulation for Baseball Pitching Machine with Rollers and Its Optimization Shinobu SAKAI*5, Yuichiro KITAGAWA, Ryo KANAI and Juhachi

kiyo5_1-masuzawa.indd

258 5) GPS 1 GPS 6) GPS DP 7) 8) 10) GPS GPS ) GPS Global Positioning System

130 Oct Radial Basis Function RBF Efficient Market Hypothesis Fama ) 4) 1 Fig. 1 Utility function. 2 Fig. 2 Value function. (1) (2)

1 Fig. 1 Extraction of motion,.,,, 4,,, 3., 1, 2. 2.,. CHLAC,. 2.1,. (256 ).,., CHLAC. CHLAC, HLAC. 2.3 (HLAC ) r,.,. HLAC. N. 2 HLAC Fig. 2

SOM SOM(Self-Organizing Maps) SOM SOM SOM SOM SOM SOM i

Q [4] 2. [3] [5] ϵ- Q Q CO CO [4] Q Q [1] i = X ln n i + C (1) n i i n n i i i n i = n X i i C exploration exploitation [4] Q Q Q ϵ 1 ϵ 3. [3] [5] [4]

平成○○年度知能システム科学専攻修士論文


Journal of Geography 116 (6) Configuration of Rapid Digital Mapping System Using Tablet PC and its Application to Obtaining Ground Truth

28 Horizontal angle correction using straight line detection in an equirectangular image

untitled

main.dvi

1 Web [2] Web [3] [4] [5], [6] [7] [8] S.W. [9] 3. MeetingShelf Web MeetingShelf MeetingShelf (1) (2) (3) (4) (5) Web MeetingShelf

Vol. 42 No. SIG 8(TOD 10) July HTML 100 Development of Authoring and Delivery System for Synchronized Contents and Experiment on High Spe

第62巻 第1号 平成24年4月/石こうを用いた木材ペレット

1 [1, 2, 3, 4, 5, 8, 9, 10, 12, 15] The Boston Public Schools system, BPS (Deferred Acceptance system, DA) (Top Trading Cycles system, TTC) cf. [13] [

149 (Newell [5]) Newell [5], [1], [1], [11] Li,Ryu, and Song [2], [11] Li,Ryu, and Song [2], [1] 1) 2) ( ) ( ) 3) T : 2 a : 3 a 1 :


(MIRU2008) HOG Histograms of Oriented Gradients (HOG)

4.1 % 7.5 %

Input image Initialize variables Loop for period of oscillation Update height map Make shade image Change property of image Output image Change time L

untitled

理学療法検査技術習得に向けた客観的臨床能力試験(OSCE)の試行

4. C i k = 2 k-means C 1 i, C 2 i 5. C i x i p [ f(θ i ; x) = (2π) p 2 Vi 1 2 exp (x µ ] i) t V 1 i (x µ i ) 2 BIC BIC = 2 log L( ˆθ i ; x i C i ) + q

IPSJ SIG Technical Report Vol.2009-DPS-141 No.20 Vol.2009-GN-73 No.20 Vol.2009-EIP-46 No /11/27 1. MIERUKEN 1 2 MIERUKEN MIERUKEN MIERUKEN: Spe

07-二村幸孝・出口大輔.indd

HASC2012corpus HASC Challenge 2010,2011 HASC2011corpus( 116, 4898), HASC2012corpus( 136, 7668) HASC2012corpus HASC2012corpus

Study on Throw Accuracy for Baseball Pitching Machine with Roller (Study of Seam of Ball and Roller) Shinobu SAKAI*5, Juhachi ODA, Kengo KAWATA and Yu

1 GPU GPGPU GPU CPU 2 GPU 2007 NVIDIA GPGPU CUDA[3] GPGPU CUDA GPGPU CUDA GPGPU GPU GPU GPU Graphics Processing Unit LSI LSI CPU ( ) DRAM GPU LSI GPU

IPSJ SIG Technical Report Vol.2014-CG-155 No /6/28 1,a) 1,2,3 1 3,4 CG An Interpolation Method of Different Flow Fields using Polar Inter

Visual Evaluation of Polka-dot Patterns Yoojin LEE and Nobuko NARUSE * Granduate School of Bunka Women's University, and * Faculty of Fashion Science,

1 I/F I/F 1 6) MobileIP 7) 8) MN: Monile Node MN AR Mobility Anchor Point(MAP) MobileIP HMIP HMIP HA-MAP MN MAP MN MAP HMIP MAP MN 2 MobileIP Mo

IPSJ SIG Technical Report Vol.2013-ARC-203 No /2/1 SMYLE OpenCL (NEDO) IT FPGA SMYLEref SMYLE OpenCL SMYLE OpenCL FPGA 1

IPSJ SIG Technical Report Vol.2010-GN-74 No /1/ , 3 Disaster Training Supporting System Based on Electronic Triage HIROAKI KOJIMA, 1 KU

xx/xx Vol. Jxx A No. xx 1 Fig. 1 PAL(Panoramic Annular Lens) PAL(Panoramic Annular Lens) PAL (2) PAL PAL 2 PAL 3 2 PAL 1 PAL 3 PAL PAL 2. 1 PAL

rank ”«‘‚“™z‡Ì GPU ‡É‡æ‡éŁÀŠñ›»

2017 (413812)

1 3DCG [2] 3DCG CG 3DCG [3] 3DCG 3 3 API 2 3DCG 3 (1) Saito [4] (a) 1920x1080 (b) 1280x720 (c) 640x360 (d) 320x G-Buffer Decaudin[5] G-Buffer D

IPSJ SIG Technical Report Vol.2014-ARC-213 No.24 Vol.2014-HPC-147 No /12/10 GPU 1,a) 1,b) 1,c) 1,d) GPU GPU Structure Of Array Array Of

日立金属技報 Vol.34

JFE.dvi

soturon.dvi


A Feasibility Study of Direct-Mapping-Type Parallel Processing Method to Solve Linear Equations in Load Flow Calculations Hiroaki Inayoshi, Non-member

Iteration 0 Iteration 1 1 Iteration 2 Iteration 3 N N N! N 1 MOPT(Merge Optimization) 3) MOPT MOP

Ł×

HP cafe HP of A A B of C C Map on N th Floor coupon A cafe coupon B Poster A Poster A Poster B Poster B Case 1 Show HP of each company on a user scree

A Study of Effective Application of CG Multimedia Contents for Help of Understandings of the Working Principles of the Internal Combustion Engine (The

IPSJ SIG Technical Report Vol.2014-GN-90 No.16 Vol.2014-CDS-9 No.16 Vol.2014-DCC-6 No /1/24 1,a) 2,b) 2,c) 1,d) QUMARION QUMARION Kinect Kinect

(a) Picking up of six components (b) Picking up of three simultaneously. components simultaneously. Fig. 2 An example of the simultaneous pickup. 6 /

06’ÓŠ¹/ŒØŒì


2010 : M DCG 3 (3DCG) 3DCG 3DCG 3DCG S

17 Proposal of an Algorithm of Image Extraction and Research on Improvement of a Man-machine Interface of Food Intake Measuring System

第 55 回自動制御連合講演会 2012 年 11 月 17 日,18 日京都大学 1K403 ( ) Interpolation for the Gas Source Detection using the Parameter Estimation in a Sensor Network S. T

1: ( 1) 3 : 1 2 4

IPSJ SIG Technical Report Vol.2009-BIO-17 No /5/26 DNA 1 1 DNA DNA DNA DNA Correcting read errors on DNA sequences determined by Pyrosequencing

1: A/B/C/D Fig. 1 Modeling Based on Difference in Agitation Method artisoc[7] A D 2017 Information Processing

( ) [1] [4] ( ) 2. [5] [6] Piano Tutor[7] [1], [2], [8], [9] Radiobaton[10] Two Finger Piano[11] Coloring-in Piano[12] ism[13] MIDI MIDI 1 Fig. 1 Syst

2. CABAC CABAC CABAC 1 1 CABAC Figure 1 Overview of CABAC 2 DCT 2 0/ /1 CABAC [3] 3. 2 値化部 コンテキスト計算部 2 値算術符号化部 CABAC CABAC

TCP/IP IEEE Bluetooth LAN TCP TCP BEC FEC M T M R M T 2. 2 [5] AODV [4]DSR [3] 1 MS 100m 5 /100m 2 MD 2 c 2009 Information Processing Society of

1

パナソニック技報

untitled

e-learning e e e e e-learning 2 Web e-leaning e 4 GP 4 e-learning e-learning e-learning e LMS LMS Internet Navigware

IPSJ SIG Technical Report Vol.2014-HCI-158 No /5/22 1,a) 2 2 3,b) Development of visualization technique expressing rainfall changing conditions

2). 3) 4) 1.2 NICTNICT DCRA Dihedral Corner Reflector micro-arraysdcra DCRA DCRA DCRA 3D DCRA PC USB PC PC ON / OFF Velleman K8055 K8055 K8055

21 e-learning Development of Real-time Learner Detection System for e-learning

エンタープライズサーチ・エンジンQ u i c k S o l u t i o n ® の開発

卒業論文2.dvi

EQUIVALENT TRANSFORMATION TECHNIQUE FOR ISLANDING DETECTION METHODS OF SYNCHRONOUS GENERATOR -REACTIVE POWER PERTURBATION METHODS USING AVR OR SVC- Ju

Table 1. Assumed performance of a water electrol ysis plant. Fig. 1. Structure of a proposed power generation system utilizing waste heat from factori

Rubin Rubin

1 OpenCL OpenCL 1 OpenCL GPU ( ) 1 OpenCL Compute Units Elements OpenCL OpenCL SPMD (Single-Program, Multiple-Data) SPMD OpenCL work-item work-group N

Vol.57 No

Journal of the Combustion Society of Japan Vol.56 No.178 (2014) FEATURE /Issues and Solutions for Engine Combustion φ-t マッ

On the Wireless Beam of Short Electric Waves. (VII) (A New Electric Wave Projector.) By S. UDA, Member (Tohoku Imperial University.) Abstract. A new e

[ ]アップル荒井

(a) 1 (b) 3. Gilbert Pernicka[2] Treibitz Schechner[3] Narasimhan [4] Kim [5] Nayar [6] [7][8][9] 2. X X X [10] [11] L L t L s L = L t + L s

IPSJ SIG Technical Report Vol.2012-IS-119 No /3/ Web A Multi-story e-picture Book with the Degree-of-interest Extraction Function

BOK body of knowledge, BOK BOK BOK 1 CC2001 computing curricula 2001 [1] BOK IT BOK 2008 ITBOK [2] social infomatics SI BOK BOK BOK WikiBOK BO

Abstract This paper concerns with a method of dynamic image cognition. Our image cognition method has two distinguished features. One is that the imag

i

Vol. 48 No. 4 Apr LAN TCP/IP LAN TCP/IP 1 PC TCP/IP 1 PC User-mode Linux 12 Development of a System to Visualize Computer Network Behavior for L


FUJII, M. and KOSAKA, M. 2. J J [7] Fig. 1 J Fig. 2: Motivation and Skill improvement Model of J Orchestra Fig. 1: Motivating factors for a

& Vol.5 No (Oct. 2015) TV 1,2,a) , Augmented TV TV AR Augmented Reality 3DCG TV Estimation of TV Screen Position and Ro

SEJulyMs更新V7

評論・社会科学 98号(P)☆/1.鰺坂

J. Jpn. Acad. Nurs. Sci. 35: (2015)

2. ICA ICA () (Blind Source Separation BBS) 2) Fig. 1 Model of Optical Topography. ( ) ICA 2.2 ICA ICA 3) n 1 1 x 1 (t) 2 x 2 (t) n x(t) 1 x(t

26.6月号inddホームページ用.indd

2 The Bulletin of Meiji University of Integrative Medicine 3, Yamashita 10 11

情報処理学会研究報告 IPSJ SIG Technical Report Vol.2015-GI-34 No /7/ % Selections of Discarding Mahjong Piece Using Neural Network Matsui

Fig. 3 Flow diagram of image processing. Black rectangle in the photo indicates the processing area (128 x 32 pixels).

1., 1 COOKPAD 2, Web.,,,,,,.,, [1]., 5.,, [2].,,.,.,, 5, [3].,,,.,, [4], 33,.,,.,,.. 2.,, 3.., 4., 5., ,. 1.,,., 2.,. 1,,

Transcription:

SOM 1 2 2 3 1 (SOM: Self-Organizing Maps) 3 SOM SOM SOM SOM GPU A Study on Visualization of Pareto Solutions by Spherical Self-Organizing Maps MASATO YOSHIMI, 1 KANAME NISHIMOTO, 2 LUYI WANG, 2 TOMOYUKI HIROYASU 3 and MITSUNORI MIKI 1 In this paper, Self-Organizing Maps (SOM) which is one of neural network systems was applied to illustrate the Pareto solution set which is derived by multi-objective optimization. Especially, this paper described that Spherical SOM is effective for illustrating the Pareto solution set. These discussions have been performed through the real world problem. At the same time, the implementation and parallel method of Spherical SOM by GPUs were also discussed. 1 Faculty of Science and Engineering, Doshisha University 2 Faculty of Engineering, Doshisha University 3 Faculty of department of life and medical science, Doshisha University 1. Computer Aided Design(CAD) Finite Element Method(FEM) 1) 2) 4) 2 3 1 2 1 c2010 Information

Fuzzy Multiple Discrimminant Analysis (FMDA) 5) (SOM) 6) SOM 3 6) SOM SOM SOM SOM SOM SOM 7) 8) SOM SOM SOM GPU 2. n k f(x) m g(x) (1) 12) { min(max) f i(x 1, x 2,..., x n) (i = 1, 2,..., k) subject to g j (x 1, x 2,..., x n ) 0 (j = 1, 2,..., m) f i (x) 1 (1) 13) x 1, x 2 I(x= (x 1, x 2,..., x n)) (a) f i (x 1 ) f i (x 2 )( i = 1, 2,..., k) x 1 x 2 (b) f i (x 1 ) < f i (x 2 )( i = 1, 2,..., k) x 1 x 2 x 1 x 2 x 1 x 2 x 0 I(x= (x 1, x 2,..., x n )) (a) x 0 x I x 0 Weak Pareto-optimal solution (b) x 0 x I x 0 Pareto-optimal solution (1) 2 k = 2 Feasible Region 3. SOM 3.1 SOM SOM SOM SOM 2 3 3.2 SOM SOM i n t m i (t) x(t) 2 c2010 Information

f (x) 2 Feasible region Weak pareto-optimal solutions Pareto-optimal solutions 1 f (x) 1 α(t) h ci(t) σ(t) SOM ( 1 ) t 0 m i (0) ( 2 ) x(t) m i (t) m i (t) x(t) ( 3 ) c x(t) = min m i (t) x(t) c ( 4 ) c d ci (2) (2) h ci(t) α(t) (3) α(t) σ(t) (4) (5) m i (t + 1) = m i (t) + h ci (t)(x(t) m i (t)) (2) { 0 (d ci > σ(t)) h ci(t) = (3) α(t) (d ci σ(t)) α(t) = α(0)(t t)/t (4) σ(t) = σ(0)(t t)/t (5) ( 5 ) t < T t t + 1 (2) t = T 4. SOM 4.1 SOM SOM SOM 3 SOM 3 SOM 4.2 4.2.1 SOM 2 2 2 4.2.2 SOM 1 6 3 c2010 Information

SOM SOM a spherical surface U development V U' 2 3D V' U' rectangular coordinates put neurons on the apices 3D 4.2.3 2 SOM 4.2.4 SOM V' 2 2 SOM 3 4 c2010 Information

h d i map e j f k m n o p q r s a b c 4.2.5 g l d 3 e tree j f k o n i a b c g l p s r not-learned neurons learn and add tree pre-learned neurons adjacent neurons e HSV (Saturation) (Value) (Hue) [0, 360] [0, 280] 5. SOM GPU 5.1 SOM SOM 1 3 SOM 1987 Carpenter SOM 1992 Speckmann SOM SIMD COKOS(COprocessor for KOhonen s Self-organizing map) COKOS 8 SOM SOM FPGA Porrmann j k p s r n f k o n i 11) Porrmann 6 SOM 128 16 16 Intel Xeon 2.80GHz CPU 350 2009 GPU SOM 10) NVIDIA GeForce GTX280 CPU 150 SOM GPU GPU SOM GPU 5.2 SOM SOM (2) SOM GPU 100 10) SOM SOM GPU 5.3 SOM SOM GPU 18) C++ NVIDIA CUDA Opteron 1210 HE(2.50GHz, Linux-2.6.26-amd64, 5 c2010 Information

G++4.1.3, -O3) 1 3 GPU 5 102 25000(20 1250) [0, 9] 4 5 1 CPU SP ( ) (GHz) (GHz) GeForce8400GS 16 0.45 0.4 GeForce GTX280 240 1.27 1.1 Tesla C1060 240 1.30 0.8 5 GPU SOM SOM 10) SOM GPU 6. SOM 4 4 (SP) GPU 10 SP 5 SOM SP 8400GS 240 SP GPU 10 SP SP 240 GPU SOM. SOM SOM SOM 3) 3) SOM SOM 6 c2010 Information

6.1 SOM, 4. SOM 2001 6)9) SOM 6.2 CO2 14) 16) (SFC) (NOX) NOX (Soot) SFC NOX Soot NOX SFC 6 2 Start Angle Exhaust gas recirculation Rate EGR RateSwirl Ratio Boost Pressure 11 2 3) NOX (Soot) SFC NOX SFC 8 6 2 f 1 NOX f 2 Soot f 3 x 1 x 2 x 3 x 4 x 5 x 6 x 7 ( ) x 8 3 8 11 6.3 SOM SOM SOM SOM SOM PAK SOM PAK 3 100000 100 1000 SOM SOM 3 4.2.4 SOM 20 20 7 c2010 Information

情報処理学会研究報告 となる ニューロン数 (個) 平面 SOM 球面 SOM 900 (30 30) 1000 (20 50) 表 3 SOM の設定 学習回数 (回) 学習率係数の初期値 100000 100000 p2: f1: 226.3017 f3: 0.0875 近傍半径の初期値 0.05 0.05 p1: f1: 458.7836 f3: 0.0963 15 15 p3: f1: 260.8934 f3: 0.0752 提案する球面 SOM によって得られた結果を図 7 に示す ここでは 90 度ずつ回転させ た様子を示している 球面上には 目的関数 f1 (燃料消費量) の値を表示している すなわ ち f1 の値が高ければ赤く 低ければ紫となる 従って 球面 SOM で解の分布を学習さ せると f1 の値が高いものが島状に配置される p5: f1: 160.1595 f3: 0.0006 p1 p4: f1: 290.4700 f3: 0.1063 図8 平面 SOM における SFC の空間の可視化 p2 p3 わかる 従って大林が平面 SOM での設計空間の分類 構造の把握を行えると示したことと p4 0 同様に 球面 SOM でも可能であると言え パレート解表示は有効であることが再確認され た9) 90 6.4 平面 SOM と球面 SOM におけるデータ配置の違い 平面 SOM が パレート解表示する上で有効であることは再確認されたが 境界において は 各競合層に配置されたデータにゆがみが生じる可能性がある そのため 図 8 において 示した平面 SOM に配置された各データが 球面 SOM ではどのように配置されているかを 検討する p5 図 8 における点 p1 から p5 までの各目的関数の値および設計変数の値を表 4 にまとめて 示す まず 図 8 における点 p2 および p4 に着目する これらの目的関数値および設計変数の 180 270 値は比較的類似しており SOM 上では 近傍に配置されることが期待される しかしなが ら 実際には平面 SOM 上では 右端に対峙して配置されている これは 類似したデータ 図 7 球面 SOM における SFC の空間の可視化 がすべて右端に偏って配置されたため 実際には近いデータにもかかわらず 距離をおいて 平面 SOM での学習結果を図 8 に示す f1 の値が高いものが左上に集まって表示され 右 配置されてしまったためであると考えられる 一方で 図 7 からもわかるように球面 SOM 上 右下 左下という順番で値が小さくなりながら配置されていることがわかる においては p2 および p4 は近傍に配置されている よって 球面 SOM によるパレート解 球面 SOM の結果も平面 SOM と同様に値を徐々に低下させながら配置されていることが 表示では境界のゆがみの影響を受けていないことが分かる 8 2010 Information

4 data f 1 f 3 f 3 x 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8 p1 458.7836 0.0004 0.0963 3.5000 0.2906 5.5781 8.2813 13.5938 10.3594 9.7656 0.5445 p2 226.3017 0.0006 0.0875 3.5688 0.2906 5.0156 14.516 3.0000 13.5313 8.8281 0.7953 p3 260.8934 0.0006 0.0752 3.4938 0.2906 5.9531 12.766 3.6566 13.5313 9.3750 0.5047 p4 290.4700 0.0004 0.1063 3.5188 0.2906 4.3125 15.391 3.7500 17.4688 8.8281 0.7625 p5 160.1595 6.2650 0.0006 3.4938 0.0000 5.9844 3.0000 17.2500-1.2500 5.7656 0.7883 8 p1 p3 8 f 1 7 8 SOM p1 p3 7 SOM 4 x 4 x 8 SOM p1 p3 SOM SOM SOM SOM SOM SOM 7. (Self-Organizing Maps, SOM) SOM SOM SOM SOM 2 3 SOM SOM SOM 1 SOM SOM SOM SOM SOM SOM SOM SOM SOM GPU SOM 1) (TOM) Vol.2, No.3 pp.14-26 (2009) 2) N700 http://trendy.nikkeibp.co.jp/article/column/20070705/1001439/?p=2 3) (TOM) Vol.1 No.1 pp.27-42, (2008) 4), Vol. 46, No. SIG 17(TOM13) pp.102-113 (2005) 5) 9 c2010 Information

Vol.20, No.6, pp.850-859, (2008) 6), : Vol.58, pp. 109-116 (2005/5) 7) Ying Xin WU SOM Vol.19, No.6, pp.611-617, (2007) 8),,, SOM Vol.8, No.1, pp.29-39 9), 14, pp.699-700 (2001) 10) HOKKE-2009 IPSJ SIG Notes pp.31-36 (2009) 11) H. Tamukoh, T. Aso, K. Horio, T. Yamakawa : Self-organizing map hardware accelerator system and its application to realtime image enlargement, Neural Networks, 2004. Proceedings. 2004 IEEE International Joint Conference on, Vol. 4, (2004) 12), 2 (1995) 13),,, :,, 17, (1998) 14), Vol.55, No.9, pp.17-22, (2001) 15) :, : Vol.55, No.9, pp.10-16, (2001) 16), :, : Vol.55, No.9, pp.41-45, (2001) 17), :, : Vol.55, No.9, pp.46-52, (2001) 18) GPU SOM HOKKE-17 IPSJ SIG Note (2009) 10 c2010 Information

( 1 ) ( 2 ) 1 CPU GPU MASATO YOSHIMI KANAME NISHIMOTO LUYI WANG TOMOYUKI HIROYASU MITSUNORI MIKI KANAME NISHIMOTO MASATO YOSHIMI LUYI WANG TOMOYUKI HIROYASU MITSUNORI MIKI