Structural SVMs 및 Pegasos 알고리즘을 이용한 한국어 개체명 인식

Size: px
Start display at page:

Download "Structural SVMs 및 Pegasos 알고리즘을 이용한 한국어 개체명 인식"

Transcription

1 Deep Learning

2 차례 현재딥러닝기술수준소개 딥러닝 딥러닝기반의자연어처리

3 Object Recognition

4 Semantic Segmentation

5 Semantic Segmentation VGGNet + Deconvolution network

6 Image Completion

7 Neural Art Artistic style transfer using CNN

8 Hand Writing by Machine LSTM RNN: Input: recurrent neural network handwriting generation demo Style:

9 Music Composition ent-neural-networks-for-folk-music-generation/

10 Image Caption Generation W t+1 Softmax Multimodal CNN GRU Image Embedding W t 한어린소녀가풀로덮인들판에서있다 건물앞에서있는한남자 구명조끼를입은한작은소녀가웃고있다 분홍색개를데리고있는한여자와한여자

11 Visual Question Answering Facebook: Visual Q&A

12 Word Analogy King Man + Woman Queen Queen King + Kings Queens Japan Korean + Hangul =?

13 Neural Machine Translation T S 777 항공편은 3 시간동안지상에 있 겠 습니다. </s> flight is on the ground for three hours </s>

14 Neural Conversation Model 세탁기가전원이안들어와요 A) 많이불편하셨겠습니다 A) 드럼이세요아니면일반세탁기세요 언제방문해줄수있나요? A) 언제방문드리면편하실까요 A) 언제방문드리면편하시겠어요 출장비가얼마죠? A) 출장점검료는 원이구요 A) 출장비가만오천원이구요

15 Abstractive Text Summarization RNN_search+input_feeding+CopyNet 로드킬로숨진친구의곁을지키는길고양이의모습이포착되었다.

16 Learning to Execute LSTM RNN

17 Learning Approximate Solutions Travelling Salesman Problem: NP-hard Pointer Network can learn approximate solutions: O(n^2)

18 One Shot Learning Learning from a few examples Matching Nets use attention and memory a(x 1,x 2 ) is a attention kernel

19 차례 현재딥러닝기술수준소개 딥러닝 딥러닝기반의자연어처리

20 Neural Networks 20

21 Deep Neural Networks Deep Neural Network = Neural Network + multiple levels of nonlinear operations. 21

22 Why Deep Neural Networks? 사람의인지과정과유사함 추상화 : 저수준의표현 고수준의표현 22

23 Why Deep Neural Networks?: Integrated Learning 기존기계학습방법론 Handcrafting features time-consuming Deep Neural Network: Feature Extractor + Classifier < 겨울학교 14 Deep Learning 자료참고 > 23

24 Why Deep Neural Networks?: Unsupervised Feature Learning 기계학습에많은학습데이터필요 소량의학습데이터 학습데이터구축비용 / 시간 대량의원시코퍼스 (unlabeled data) Semi-supervised, Unsupervised Deep Neural Network Pre-training 방법을통해대량의원시코퍼스에서자질학습 Restricted Boltzmann Machines (RBM) Stacked Autoencoder, Stacked Denosing Autoencoder Word Embedding (for NLP) 24

25 DNN Difficulties Now 학습이잘안됨 Unsupervised Pre-training Back-propagation 알고리즘 X 많은계산이필요함 하드웨어 /GPU 발전 Many parameters Over-fitting 문제 Pre-training, Drop-out, 25

26 Deep Belief Network [Hinton06] Key idea Pre-train layers with an unsupervised learning algorithm in phases Then, fine-tune the whol e network by supervised learning DBN are stacks of Restricted Boltzmann Machines (RBM) 26

27 Restricted Boltzmann Machine A Restricted Boltzmann ma chine (RBM) is a generative stochastic neural network that can learn a probability distribution over its set of inputs Major applications Dimensionality reduction Topic modeling, 27

28 Training DBN: Pre-Training 1. Layer-wise greedy unsupervised pre-training Train layers in phase from the bottom layer 28

29 Training DBN: Fine-Tuning 2. Supervised fine-tuning for the classification task 29

30 The Back-Propagation Algorithm

31 Autoencoder Autoencoder is an NN whose desired output is the same as the input To learn a compressed representation (encoding) for a set of data. Find weight vectors A and B that minimize: Σ i (y i -x i ) 2 < 겨울학교 14 Deep Learning 자료참고 > 31

32 Stacked Autoencoders After training, the hidden node extracts features from the input nodes Stacking autoencoders constructs a deep network < 겨울학교 14 Deep Learning 자료참고 > 32

33 Dropout (Hinton12) In training, randomly dropout hidden units with probability p. < 겨울학교 14 Deep Learning 자료참고 > 33

34 Non-linearity (Activation Function) 34

35 Convolutional Neural Network (LeCun98) Convolutional NN Convolution Layer Sparse Connectivity Shared Weights Multiple feature maps Sub-sampling Layer Ex. LeNet Average/max pooling NxN 1 Multiple feature maps

36 CNN Architectures

37 CNN for Audio

38 Recurrent Neural Network Recurrent property dynamical system over time

39 Bidirectional RNN Exploit future context as well as past

40 Long Short-Term Memory RNN LSTM can preserve gradient information

41 차례 현재딥러닝기술수준소개 딥러닝 딥러닝기반의자연어처리

42 텍스트의표현방식 One-hot representation (or symbolic) Ex. [ ] Dimensionality 50K (PTB) 500K (big vocab) 3M (Google 1T) Problem Motel [ ] AND Hotel [ ] = 0 Continuous representation Latent Semantic Analysis, Random projection Latent Dirichlet Allocation, HMM clustering Neural Word Embedding Dense vector By adding supervision from other tasks improve the representation

43 Neural Network Language Model (Bengio00,03) Idea A word and its context is a positive training sample A random word in that sam e context negative trainin g sample Score(positive) > Score(neg.) Training complexity is high Hidden layer output Softmax in the output layer Hierarchical softmax Negative sampling Ranking(hinge loss) LT: V *d, Input(one hot): V *1 LT T I Shared weights = Word embedding Input Dim: 1 Dim: 2 Dim: 3 Dim: 4 Dim: 5 1 (boy) (girl)

44 한국어 Word Embedding: NNLM

45 전이기반의한국어의존구문분석 : Forward Transition-based(Arc-Eager): O(N) 예 : CJ 그룹이 1 대한통운 2 인수계약을 3 체결했다 4 [root], [CJ 그룹이 1 대한통운 2 ], {} 1: Shift [root CJ 그룹이 1 ], [ 대한통운 2 인수계약을 3 ], {} 2: Shift [root CJ 그룹이 1 대한통운 2 ], [ 인수계약을 3 체결했다 4 ], {} 3: Left-arc(NP_MOD) [root CJ 그룹이 1 ], [2 인수계약을 3 체결했다 4 ], {( 인수계약을 3 대한통운 2 )} 4: Shift [root CJ 그룹이 1 2 인수계약을 3 ], [ 체결했다 4 ], {( 인수계약을 3 대한통운 2 )} 5: Left-arc(NP_OBJ) [root CJ 그룹이 1 ], [3 체결했다 4 ], {( 체결했다 4 인수계약을 3 ), } 6: Left-arc(NP_SUB) [root], [(1,3) 체결했다 4 ], {( 체결했다 4 CJ 그룹이 1 ), } 7: Right-arc(VP) [root 4 (1,3) 체결했다 4 ], [], {(root 체결했다 4 ), }

46 딥러닝기반한국어의존구문분석 ( 한글및한국어 14) Transition-based + Backward O(N) 세종코퍼스 의존구문변환 보조용언 / 의사보조용언후처리 Deep Learning 기반 ReLU(> Sigmoid) + Dropout Korean Word Embedding NNLM, Ranking(hinge, logit) Word2Vec Feature Embedding POS (stack + buffer) 자동분석 ( 오류포함 ) Dependency Label (stack) Distance information Valency information Mutual Information 대용량코퍼스 자동구문분석 LT 1 LT N Input Word S[w t-2 w t-1 ] B[w t ] Word Lookup Table M 1 x M 2 x Linear ReLU Linear concat Input Feature f 1 f 2 f 3 LT 1 f 4 Feature Lookup Table LT D h #output

47 한국어의존구문분석실험결과 기존연구 : UAS 85~88% Structural SVM 기반성능 : UAS=89.99% LAS=87.74% Pre-training > no Pre. Dropout > no Dropout ReLU > Sigmoid MI feat. > no MI feat. Word Embedding 성능순위 1. NNLM 2. Ranking(logit loss) 3. Word2vec 4. Ranking(hinge loss)

48 LSTM RNN + CRF LSTM-CRF 제안 y(t+1) y(t ) y(t+1) y(t+1) y(t ) y(t+1) h(t-1) h(t ) h(t+1) x(t-1) x(t ) x(t+1) x(t-1) x(t ) x(t+1) y(t+1) y(t ) y(t+1) f (t ) h(t-1) h(t ) h(t+1) i (t ) o(t ) x(t-1) x(t ) x(t+1) x(t ) C(t) h(t )

49 영어개체명인식 (KCC 15, Journal submitted) 영어개체명인식 (CoNLL03 data set) F1(dev) F1(test) SENNA (Collobert) Structural SVM (baseline + Word embedding feature) FFNN (Sigm + Dropout + Word embedding) RNN (Sigm + Dropout + Word embedding) LSTM RNN (Sigm + Dropout + Word embedding) GRU RNN (Sigm + Dropout + Word embedding) CNN+CRF (Sigm + Dropout + Word embedding) RNN+CRF (Sigm + Dropout + Word embedding) LSTM+CRF (Sigm + Dropout + Word embedding) GRU+CRF (Sigm + Dropout + Word embedding)

50 한국어감성분석 CNN Mobile data Train: 4543, Test: 500 EMNLP14 모델 (CNN) 적용 Matlab 으로구현 Word embedding: 한국어 10 만단어 + 도메인특화 1420 단어 Data set Model Accuracy Mobile Train: 4543 Test: 500 SVM (word feature) CNN(relu,kernel3,hid50)+Word embedding (word feature) 91.20

51 LSTM RNN 기반한국어감성분석 LSTM RNN-based encoding Sentence embedding 입력 Fully connected NN 출력 GRU encoding 도유사함 h(1) h(2 ) h(t) y x(1) x(2 ) x(t) Data set Model Accuracy Mobile Train: 4543 Test: 500 SVM (word feature) CNN(relu,kernel3,hid50)+Word embedding (word feature) GRU encoding + Fully connected NN LSTM RNN encoding + Fully connected NN 90.93

52 Recurrent NN Encoder Decoder for Statistical Machine Translation (EMNLP14)

53 Sequence to Sequence Learning with Neural Networks (NIPS14 Google) Source Voc.: 160,000 Target Voc.: 80,000 Deep LSTMs with 4 layers Train: 7.5 epochs (12M sentences, 10 days with 8- GPU machine)

54 Neural MT by Jointly Learning to Align and Translate (ICLR15) GRU RNN + Alignment Encoding GRU RNN Decoding Vocab: 30,000 (src, tgt) Train: 5 days

55 J-to-E Neural MT (WAT) 1/2 ASPEC-JE data Neural MT (RNN-search) GRU RNN + Alignment Encoding GRU RNN Decoding Vocab size: 20,000 (src, tgt) BLEU(test): (beam=10) WAT14(Juman): PBMT=18.45, HPBMT=18.72, NAIST(1 위,forest-tostring)=23.29

56 J-to-E Neural MT 실험 (WAT) 2/2 最後 /ncc:0 に /ps:1,/sl:2 将来 /nca:3 展望 /ncs:4 に /ps:5 つい /vc:6 て /pj:7 記述 /ncs:8 </s>:9 the/dt:0 future/jj:1 view/nn:2 is/vbz:3 described/vbn:4./.:5 </s>:6 食物 /ncc:0 アレルギー /ncc:1 は /pc:2 アナフィラキシー /ncc:3 の /ps:4 主要 /dc:5 な /vx:6 原因 /ncs:7 抗原 /ncc:8 の /ps:9 一 /nn:10 つ /xnn:11 で /vx:12 ある /vd:13 /op:14 </s>:15 the/dt:0 food/nn:1 allergy/nn:2 is/vbz:3 one/cd:4 of/in:5 the/dt:6 main/jj:7 causal/jj:8 antigen/nn:9 of/in:10 the/dt:11 anaphylaxis/nn:12./.:13 </s>:14

57 Neural Conversation Model 세탁기가전원이안들어와요 A) 많이불편하셨겠습니다 A) 드럼이세요아니면일반세탁기세요 언제방문해줄수있나요? A) 언제방문드리면편하실까요 A) 언제방문드리면편하시겠어요 출장비가얼마죠? A) 출장점검료는 원이구요 A) 출장비가만오천원이구요

58 Abstractive Text Summarization RNN_search+input_feeding+CopyNet 로드킬로숨진친구의곁을지키는길고양이의모습이포착되었다.

59 Learning to Execute LSTM RNN

60 Learning Approximate Solutions Travelling Salesman Problem: NP-hard Pointer Network can learn approximate solutions: O(n^2)

61 End-to-End Neural Speech Recognition (15)

62 Neural Image Caption Generator (14)

63 Korean Image Caption Generation W t+1 Softmax Multimodal CNN GRU Image Embedding W t 한어린소녀가풀로덮인들판에서있다 건물앞에서있는한남자 구명조끼를입은한작은소녀가웃고있다 분홍색개를데리고있는한여자와한여자

RNN & NLP Application

RNN & NLP Application RNN & NLP Application 강원대학교 IT 대학 이창기 차례 RNN NLP application Recurrent Neural Network Recurrent property dynamical system over time Bidirectional RNN Exploit future context as well as past Long Short-Term

More information

PowerPoint Presentation

PowerPoint Presentation Dependency Parser 자연언어처리 Probabilistic CFG (PCFG) - CFG - PCFG with saw with saw astronomers ears saw stars telescope astronomers ears saw stars telescope PCFG example Repeated work Parsing PCFG: CKY CKY

More information

Structural SVMs 및 Pegasos 알고리즘을 이용한 한국어 개체명 인식

Structural SVMs 및 Pegasos 알고리즘을 이용한 한국어 개체명 인식 딥러닝 기반의 자연어처리 기술 강원대학교 IT대학 이창기 차례 자연어처리소개 딥러닝소개 딥러닝기반의자연어처리 Classification Problem Sequence Labeling Problem Sequence-to-Sequence Learning Pointer Network Machine Reading Comprehension 자연어처리 자연언어 인공언어에대응되는개념

More information

딥러닝NLP응용_이창기

딥러닝NLP응용_이창기 딥러닝과 자연어처리 응용 강원대학교 IT대학 이창기 차례 딥러닝최신기술소개 딥러닝기반의자연어처리 Classification Problem Sequence Labeling Problem Sequence-to-Sequence Learning Pointer Network Recurrent Neural Network Many NLP problems can be viewed

More information

<4D6963726F736F667420576F7264202D20B1E2C8B9BDC3B8AEC1EE2DC0E5C7F5>

<4D6963726F736F667420576F7264202D20B1E2C8B9BDC3B8AEC1EE2DC0E5C7F5> 주간기술동향 2016. 5.18. 컴퓨터 비전과 인공지능 장혁 한국전자통신연구원 선임연구원 최근 많은 관심을 받고 있는 인공지능(Artificial Intelligence: AI)의 성과는 뇌의 작동 방식과 유사한 딥 러닝의 등장에 기인한 바가 크다. 이미 미국과 유럽 등 AI 선도국에서는 인공지능 연구에서 인간 뇌 이해의 중요성을 인식하고 관련 대형 프로젝트들을

More information

Delving Deeper into Convolutional Networks for Learning Video Representations - Nicolas Ballas, Li Yao, Chris Pal, Aaron Courville arXiv:

Delving Deeper into Convolutional Networks for Learning Video Representations  -   Nicolas Ballas, Li Yao, Chris Pal, Aaron Courville  arXiv: Delving Deeper into Convolutional Networks for Learning Video Representations Nicolas Ballas, Li Yao, Chris Pal, Aaron Courville arxiv: 1511.06432 Il Gu Yi DeepLAB in Modu Labs. June 13, 2016 Il Gu Yi

More information

2 : (Seungsoo Lee et al.: Generating a Reflectance Image from a Low-Light Image Using Convolutional Neural Network) (Regular Paper) 24 4, (JBE

2 : (Seungsoo Lee et al.: Generating a Reflectance Image from a Low-Light Image Using Convolutional Neural Network) (Regular Paper) 24 4, (JBE 2: (Seungsoo Lee et al.: Generating a Reflectance Image from a Low-Light Image Using Convolutional Neural Network) (Regular Paper) 24 4, 2019 7 (JBE Vol. 24, No. 4, July 2019) https://doi.org/10.5909/jbe.2019.24.4.623

More information

PowerPoint 프레젠테이션

PowerPoint 프레젠테이션 Visual Search At SK-Planet sk-planet Machine Intelligence Lab. 나상일 1. 개발배경 2. 첫접근방법 3. 개선된방법 A. Visual recognition technology B. Guided search C. Retrieval system 개발배경 개발배경 상품검색을좀더쉽게 Key-word 트렌치코트버튺벨트

More information

Multi-pass Sieve를 이용한 한국어 상호참조해결 반-자동 태깅 도구

Multi-pass Sieve를 이용한 한국어 상호참조해결 반-자동 태깅 도구 Siamese Neural Network 박천음 강원대학교 Intelligent Software Lab. Intelligent Software Lab. Intro. S2Net Siamese Neural Network(S2Net) 입력 text 들을 concept vector 로표현하기위함에기반 즉, similarity 를위해가중치가부여된 vector 로표현

More information

4 : (Hyo-Jin Cho et al.: Audio High-Band Coding based on Autoencoder with Side Information) (Special Paper) 24 3, (JBE Vol. 24, No. 3, May 2019

4 : (Hyo-Jin Cho et al.: Audio High-Band Coding based on Autoencoder with Side Information) (Special Paper) 24 3, (JBE Vol. 24, No. 3, May 2019 4 : (Hyo-Jin Cho et al.: Audio High-Band Coding based on Autoencoder with Side Information) (Special Paper) 24 3, 2019 5 (JBE Vol. 24, No. 3, May 2019) https://doi.org/10.5909/jbe.2019.24.3.387 ISSN 2287-9137

More information

DIY 챗봇 - LangCon

DIY 챗봇 - LangCon without Chatbot Builder & Deep Learning bage79@gmail.com Chatbot Builder (=Dialogue Manager),. We need different chatbot builders for various chatbot services. Chatbot builders can t call some external

More information

_KrlGF발표자료_AI

_KrlGF발표자료_AI AI 의과거와현재그리고내일 AI is the New Electricity 2017.09.15 AI! 2 Near Future of Super Intelligence? *source l http://www.motherjones.com/media/2013/05/robots-artificial-intelligence-jobs-automation 3 4 I think

More information

(JBE Vol. 24, No. 1, January 2019) (Special Paper) 24 1, (JBE Vol. 24, No. 1, January 2019) ISSN 2287-

(JBE Vol. 24, No. 1, January 2019) (Special Paper) 24 1, (JBE Vol. 24, No. 1, January 2019)   ISSN 2287- (Special Paper) 24 1 2019 1 (JBE Vol. 24 No. 1 January 2019) https//doi.org/10.5909/jbe.2019.24.1.58 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a) a) a) b) c) d) A Study on Named Entity Recognition

More information

김기남_ATDC2016_160620_[키노트].key

김기남_ATDC2016_160620_[키노트].key metatron Enterprise Big Data SKT Metatron/Big Data Big Data Big Data... metatron Ready to Enterprise Big Data Big Data Big Data Big Data?? Data Raw. CRM SCM MES TCO Data & Store & Processing Computational

More information

R을 이용한 텍스트 감정분석

R을 이용한 텍스트 감정분석 R Data Analyst / ( ) / kim@mindscale.kr (kim@mindscale.kr) / ( ) ( ) Analytic Director R ( ) / / 3/45 4/45 R? 1. : / 2. : ggplot2 / Web 3. : slidify 4. : 5. Matlab / Python -> R Interactive Plots. 5/45

More information

(JBE Vol. 23, No. 2, March 2018) (Special Paper) 23 2, (JBE Vol. 23, No. 2, March 2018) ISSN

(JBE Vol. 23, No. 2, March 2018) (Special Paper) 23 2, (JBE Vol. 23, No. 2, March 2018)   ISSN (Special Paper) 23 2, 2018 3 (JBE Vol. 23, No. 2, March 2018) https://doi.org/10.5909/jbe.2018.23.2.186 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a), a) Robust Online Object Tracking via Convolutional

More information

THE JOURNAL OF KOREAN INSTITUTE OF ELECTROMAGNETIC ENGINEERING AND SCIENCE Jul.; 29(7),

THE JOURNAL OF KOREAN INSTITUTE OF ELECTROMAGNETIC ENGINEERING AND SCIENCE Jul.; 29(7), THE JOURNAL OF KOREAN INSTITUTE OF ELECTROMAGNETIC ENGINEERING AND SCIENCE. 2018 Jul.; 29(7), 550 559. http://dx.doi.org/10.5515/kjkiees.2018.29.7.550 ISSN 1226-3133 (Print) ISSN 2288-226X (Online) Human

More information

딥러닝 첫걸음

딥러닝 첫걸음 딥러닝첫걸음 4. 신경망과분류 (MultiClass) 다범주분류신경망 Categorization( 분류 ): 예측대상 = 범주 이진분류 : 예측대상범주가 2 가지인경우 출력층 node 1 개다층신경망분석 (3 장의내용 ) 다범주분류 : 예측대상범주가 3 가지이상인경우 출력층 node 2 개이상다층신경망분석 비용함수 : Softmax 함수사용 다범주분류신경망

More information

(JBE Vol. 24, No. 2, March 2019) (Special Paper) 24 2, (JBE Vol. 24, No. 2, March 2019) ISSN

(JBE Vol. 24, No. 2, March 2019) (Special Paper) 24 2, (JBE Vol. 24, No. 2, March 2019)   ISSN (Special Paper) 24 2, 2019 3 (JBE Vol. 24, No. 2, March 2019) https://doi.org/10.5909/jbe.2019.24.2.234 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) SIFT a), a), a), a) SIFT Image Feature Extraction

More information

01 AI Definition 02 Deep Learning Theory - Linear Regression - Cost Function - Gradient Descendent - Logistic Regression - Activation Function - Conce

01 AI Definition 02 Deep Learning Theory - Linear Regression - Cost Function - Gradient Descendent - Logistic Regression - Activation Function - Conce Artificial Intelligence for Deep Learning 01 AI Definition 02 Deep Learning Theory - Linear Regression - Cost Function - Gradient Descendent - Logistic Regression - Activation Function - Concept of Neural

More information

2 : (EunJu Lee et al.: Speed-limit Sign Recognition Using Convolutional Neural Network Based on Random Forest). (Advanced Driver Assistant System, ADA

2 : (EunJu Lee et al.: Speed-limit Sign Recognition Using Convolutional Neural Network Based on Random Forest). (Advanced Driver Assistant System, ADA (JBE Vol. 20, No. 6, November 2015) (Regular Paper) 20 6, 2015 11 (JBE Vol. 20, No. 6, November 2015) http://dx.doi.org/10.5909/jbe.2015.20.6.938 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a), a),

More information

Slide 1

Slide 1 Clock Jitter Effect for Testing Data Converters Jin-Soo Ko Teradyne 2007. 6. 29. 1 Contents Noise Sources of Testing Converter Calculation of SNR with Clock Jitter Minimum Clock Jitter for Testing N bit

More information

Microsoft PowerPoint - 실습소개와 AI_ML_DL_배포용.pptx

Microsoft PowerPoint - 실습소개와 AI_ML_DL_배포용.pptx 실습강의개요와인공지능, 기계학습, 신경망 < 인공지능입문 > 강의 허민오 Biointelligence Laboratory School of Computer Science and Engineering Seoul National University 실습강의개요 노트북을꼭지참해야하는강좌 신경망소개 (2 주, 허민오 ) Python ( 프로그래밍언어 ) (2주, 김준호

More information

Manufacturing6

Manufacturing6 σ6 Six Sigma, it makes Better & Competitive - - 200138 : KOREA SiGMA MANAGEMENT C G Page 2 Function Method Measurement ( / Input Input : Man / Machine Man Machine Machine Man / Measurement Man Measurement

More information

Interactive Transcribed Dialog Data Normalization

Interactive Transcribed Dialog Data Normalization [ Introduction ] Deep Learning 정상근 2015-11-03 Applications WHAT MACHINES CAN DO? (2015) Video Understanding (Real-time Genre Detection) Google Image Understanding Google http://techcrunch.com/2014/11/18/new-google-research-project-can-auto-caption-complex-images/

More information

사회통계포럼

사회통계포럼 wcjang@snu.ac.kr Acknowledgements Dr. Roger Peng Coursera course. https://github.com/rdpeng/courses Creative Commons by Attribution /. 10 : SNS (twitter, facebook), (functional data) : (, ),, /Data Science

More information

4 : CNN (Sangwon Suh et al.: Dual CNN Structured Sound Event Detection Algorithm Based on Real Life Acoustic Dataset) (Regular Paper) 23 6, (J

4 : CNN (Sangwon Suh et al.: Dual CNN Structured Sound Event Detection Algorithm Based on Real Life Acoustic Dataset) (Regular Paper) 23 6, (J (Regular Paper) 23 6, 2018 11 (JBE Vol. 23, No. 6, November 2018) https://doi.org/10.5909/jbe.2018.23.6.855 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) CNN a), a), a), a), a) Dual CNN Structured Sound

More information

(JBE Vol. 23, No. 2, March 2018) (Special Paper) 23 2, (JBE Vol. 23, No. 2, March 2018) ISSN

(JBE Vol. 23, No. 2, March 2018) (Special Paper) 23 2, (JBE Vol. 23, No. 2, March 2018)   ISSN (Special Paper) 23 2, 2018 3 (JBE Vol. 23, No. 2, March 2018) https://doi.org/10.5909/jbe.2018.23.2.246 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) CNN a), a), a) CNN-Based Hand Gesture Recognition

More information

<4D F736F F F696E74202D F ABFACB1B8C8B85FBEF0BEEEC3B3B8AEBFCDB1E2B0E8B9F8BFAAC7F6C8B228C1F6C3A2C1F829>

<4D F736F F F696E74202D F ABFACB1B8C8B85FBEF0BEEEC3B3B8AEBFCDB1E2B0E8B9F8BFAAC7F6C8B228C1F6C3A2C1F829> Ebiz 연구회 2017 9 21 정의용 FrankJeong@systrangroupcom SYSTRAN History & Technology Natural Language Processing Machine Translation History MT Technique Neural Network Neural Machine Translation Data Landscape

More information

본문01

본문01 Ⅱ 논술 지도의 방법과 실제 2. 읽기에서 논술까지 의 개발 배경 읽기에서 논술까지 자료집 개발의 본래 목적은 초 중 고교 학교 평가에서 서술형 평가 비중이 2005 학년도 30%, 2006학년도 40%, 2007학년도 50%로 확대 되고, 2008학년도부터 대학 입시에서 논술 비중이 커지면서 논술 교육은 학교가 책임진다. 는 풍토 조성으로 공교육의 신뢰성과

More information

02본문

02본문 46 특집 : 딥러닝기반방송미디어기술 특집 딥러닝기반방송미디어기술 딥러닝기반의음성 / 오디오기술 Speech/Audio Processing based on Deep Learning 이영한 / KETI Ⅰ. 서론 인간의두뇌를모델링하는뉴럴네트워크연구는 1940 년대신경세포의모델링부터시작하여현재까지다양한기술이축적되어왔다. 특히 backpropagation 이제안된이후에

More information

PowerPoint 프레젠테이션

PowerPoint 프레젠테이션 1 2 3 3-1 3-2 3-3 Large-scale data 개요기계학습기반자료분석기술교통분석에기계학습적용사례 1 CNN 을활용한대중교통수요예측 2 RNN 을활용한공로통행속도예측 3 DQN 을이용한최적교통신호제어 4 시연 Large-Scale Data 기술요소 전수자료 이력자료누적 ( 자료를지우지않음 ) Hadoop HDFS MapReduce 병렬 DBMS

More information

김경재 안현철 지능정보연구제 17 권제 4 호 2011 년 12 월

김경재 안현철 지능정보연구제 17 권제 4 호 2011 년 12 월 지능정보연구제 17 권제 4 호 2011 년 12 월 (pp.241~254) Support vector machines(svm),, CRM. SVM,,., SVM,,.,,. SVM, SVM. SVM.. * 2009() (NRF-2009-327- B00212). 지능정보연구제 17 권제 4 호 2011 년 12 월 김경재 안현철 지능정보연구제 17 권제 4 호

More information

<313120C0AFC0FCC0DA5FBECBB0EDB8AEC1F2C0BB5FC0CCBFEBC7D15FB1E8C0BAC5C25FBCF6C1A42E687770>

<313120C0AFC0FCC0DA5FBECBB0EDB8AEC1F2C0BB5FC0CCBFEBC7D15FB1E8C0BAC5C25FBCF6C1A42E687770> 한국지능시스템학회 논문지 2010, Vol. 20, No. 3, pp. 375-379 유전자 알고리즘을 이용한 강인한 Support vector machine 설계 Design of Robust Support Vector Machine Using Genetic Algorithm 이희성 홍성준 이병윤 김은태 * Heesung Lee, Sungjun Hong,

More information

PowerPoint 프레젠테이션

PowerPoint 프레젠테이션 I. 문서표준 1. 문서일반 (HY중고딕 11pt) 1-1. 파일명명체계 1-2. 문서등록정보 2. 표지표준 3. 개정이력표준 4. 목차표준 4-1. 목차슬라이드구성 4-2. 간지슬라이드구성 5. 일반표준 5-1. 번호매기기구성 5-2. 텍스트박스구성 5-3. 테이블구성 5-4. 칼라테이블구성 6. 적용예제 Machine Learning Credit Scoring

More information

Ch 1 머신러닝 개요.pptx

Ch 1 머신러닝 개요.pptx Chapter 1. < > :,, 2017. Slides Prepared by,, Biointelligence Laboratory School of Computer Science and Engineering Seoul National University 1.1 3 1.2... 7 1.3 10 1.4 16 1.5 35 2 1 1.1 n,, n n Artificial

More information

3 Gas Champion : MBB : IBM BCS PO : 2 BBc : : /45

3 Gas Champion : MBB : IBM BCS PO : 2 BBc : : /45 3 Gas Champion : MBB : IBM BCS PO : 2 BBc : : 20049 0/45 Define ~ Analyze Define VOB KBI R 250 O 2 2.2% CBR Gas Dome 1290 CTQ KCI VOC Measure Process Data USL Target LSL Mean Sample N StDev (Within) StDev

More information

PowerPoint Presentation

PowerPoint Presentation 기계학습을통한 시계열데이터분석및 금융시장예측응용 울산과학기술원 전기전자컴퓨터공학부최재식 얼굴인식 Facebook 의얼굴인식기 (DeepFace) 가사람과비슷한인식성능을보임 문제 : 사진에서연애인의이름을맞추기 사람의인식율 : 97.5% vs DeepFace 의인식률 : 97.35% (2014 년 3 월 ) 물체인식 ImageNet (http://image-net.org):

More information

Buy one get one with discount promotional strategy

Buy one get one with discount promotional strategy Buy one get one with discount Promotional Strategy Kyong-Kuk Kim, Chi-Ghun Lee and Sunggyun Park ISysE Department, FEG 002079 Contents Introduction Literature Review Model Solution Further research 2 ISysE

More information

High Resolution Disparity Map Generation Using TOF Depth Camera In this paper, we propose a high-resolution disparity map generation method using a lo

High Resolution Disparity Map Generation Using TOF Depth Camera In this paper, we propose a high-resolution disparity map generation method using a lo High Resolution Disparity Map Generation Using TOF Depth Camera In this paper, we propose a high-resolution disparity map generation method using a low-resolution Time-Of- Flight (TOF) depth camera and

More information

KCC2011 우수발표논문 휴먼오피니언자동분류시스템구현을위한비결정오피니언형용사구문에대한연구 1) Study on Domain-dependent Keywords Co-occurring with the Adjectives of Non-deterministic Opinion

KCC2011 우수발표논문 휴먼오피니언자동분류시스템구현을위한비결정오피니언형용사구문에대한연구 1) Study on Domain-dependent Keywords Co-occurring with the Adjectives of Non-deterministic Opinion KCC2011 우수발표논문 휴먼오피니언자동분류시스템구현을위한비결정오피니언형용사구문에대한연구 1) Study on Domain-dependent Keywords Co-occurring with the Adjectives of Non-deterministic Opinion 요약 본연구에서는, 웹문서로부터특정상품에대한의견문장을분석하는오피니언마이닝 (Opinion

More information

종합설계 I (Xcode and Source Control )

종합설계 I  (Xcode and Source Control ) Sogang University: Dept of Computer Science 구명완교수 서강대학교컴퓨터공학과 Email: mwkoo9@gmail.com 내용 개념도 시맨틱해석기 대화상태추적기 대화관리기 자연언어생성기 결론 개념도 음성인식 음성대화 음성대화솔류션사례 음성대화서비스 음성대화로봇 Apple Siri Google MS Cortana Amazon Echo

More information

Page 2 of 5 아니다 means to not be, and is therefore the opposite of 이다. While English simply turns words like to be or to exist negative by adding not,

Page 2 of 5 아니다 means to not be, and is therefore the opposite of 이다. While English simply turns words like to be or to exist negative by adding not, Page 1 of 5 Learn Korean Ep. 4: To be and To exist Of course to be and to exist are different verbs, but they re often confused by beginning students when learning Korean. In English we sometimes use the

More information

Microsoft PowerPoint - AC3.pptx

Microsoft PowerPoint - AC3.pptx Chapter 3 Block Diagrams and Signal Flow Graphs Automatic Control Systems, 9th Edition Farid Golnaraghi, Simon Fraser University Benjamin C. Kuo, University of Illinois 1 Introduction In this chapter,

More information

<4D6963726F736F667420576F7264202D20C3D6BDC52049435420C0CCBDB4202D20BAB9BBE7BABB>

<4D6963726F736F667420576F7264202D20C3D6BDC52049435420C0CCBDB4202D20BAB9BBE7BABB> 주간기술동향 2016. 2. 24. 최신 ICT 이슈 인공지능 바둑 프로그램 경쟁, 구글이 페이스북에 리드 * 바둑은 경우의 수가 많아 컴퓨터가 인간을 넘어서기 어려움을 보여주는 사례로 꼽혀 왔 으며, 바로 그런 이유로 인공지능 개발에 매진하는 구글과 페이스북은 바둑 프로그램 개 발 경쟁을 벌여 왔으며, 프로 9 단에 도전장을 낸 구글이 일단 한발 앞서 가는

More information

PowerPoint Presentation

PowerPoint Presentation 컴퓨터비전 및 패턴인식 연구회 2009.2.12 Support Vector Machines http://cespc1.kumoh.ac.kr/~nonezero/svm ws cvpr.pdf 금오공과대학교 컴퓨터공학부 고재필 1 Contents Introduction Optimal Hyperplane Soft-Margin SVM Nonlinear SVM with Kernel

More information

방송공학회논문지 제18권 제2호

방송공학회논문지 제18권 제2호 방송공학회논문지 제 20권 6호 (2015년 11월) 특집논문 : 2015년 하계학술대회 좌장추천 우수논문 프레넬 회절을 이용한 디지털 홀로그램 암호화 알고리즘 새로운 광적응 효과 모델을 이용한 정교한 영상 화질 측정 민방위 경보 방송에 대한 정보 수용자 인식 연구 UHDTV 방송을 위한 공간 변조 다중 안테나 시스템 수신 성능 분석 홍보동영상 제작 서비스를

More information

2 : (Juhyeok Mun et al.: Visual Object Tracking by Using Multiple Random Walkers) (Special Paper) 21 6, (JBE Vol. 21, No. 6, November 2016) ht

2 : (Juhyeok Mun et al.: Visual Object Tracking by Using Multiple Random Walkers) (Special Paper) 21 6, (JBE Vol. 21, No. 6, November 2016) ht (Special Paper) 21 6, 2016 11 (JBE Vol. 21, No. 6, November 2016) http://dx.doi.org/10.5909/jbe.2016.21.6.913 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a), a), a) Visual Object Tracking by Using Multiple

More information

<4D6963726F736F667420576F7264202D20C3D6BDC52049435420C0CCBDB4202D20BAB9BBE7BABB>

<4D6963726F736F667420576F7264202D20C3D6BDC52049435420C0CCBDB4202D20BAB9BBE7BABB> 최신 ICT 이슈 최신 ICT 이슈 알파고의 심층강화학습을 뒷받침한 H/W 와 S/W 환경의 진화 * 알파고의 놀라운 점은 바둑의 기본규칙조차 입력하지 않았지만 승리 방식을 스스로 알아 냈다는 것이며, 알파고의 핵심기술인 심층강화학습이 급속도로 발전한 배경에는 하드웨 어의 진화와 함께 오픈소스화를 통해 발전하는 AI 관련 소프트웨어들이 자리하고 있음 2014

More information

산선생의 집입니다. 환영해요

산선생의 집입니다. 환영해요 Biped Walking Robot Biped Walking Robot Simulation Program Down(Visual Studio 6.0 ) ). Version.,. Biped Walking Robot - Project Degree of Freedom : 12(,,, 12) :,, : Link. Kinematics. 1. Z (~ Diablo Set

More information

Software Requirrment Analysis를 위한 정보 검색 기술의 응용

Software Requirrment Analysis를 위한 정보 검색 기술의 응용 EPG 정보 검색을 위한 예제 기반 자연어 대화 시스템 김석환 * 이청재 정상근 이근배 포항공과대학교 컴퓨터공학과 지능소프트웨어연구실 {megaup, lcj80, hugman, gblee}@postech.ac.kr An Example-Based Natural Language System for EPG Information Access Seokhwan Kim

More information

Electronics and Telecommunications Trends 인공지능을이용한 3D 콘텐츠기술동향및향후전망 Recent Trends and Prospects of 3D Content Using Artificial Intelligence Technology

Electronics and Telecommunications Trends 인공지능을이용한 3D 콘텐츠기술동향및향후전망 Recent Trends and Prospects of 3D Content Using Artificial Intelligence Technology Electronics and Telecommunications Trends 인공지능을이용한 3D 콘텐츠기술동향및향후전망 Recent Trends and Prospects of 3D Content Using Artificial Intelligence Technology 이승욱 (S.W. Lee, tajinet@etri.re.kr) 황본우 (B.W. Hwang,

More information

PowerPoint 프레젠테이션

PowerPoint 프레젠테이션 CRM Fair 2004 Spring Copyright 2004 DaumSoft All rights reserved. INDEX Copyright 2004 DaumSoft All rights reserved. Copyright 2004 DaumSoft All rights reserved. Copyright 2004 DaumSoft All rights reserved.

More information

Journal of Educational Innovation Research 2018, Vol. 28, No. 4, pp DOI: * A S

Journal of Educational Innovation Research 2018, Vol. 28, No. 4, pp DOI:   * A S Journal of Educational Innovation Research 2018, Vol. 28, No. 4, pp.461-487 DOI: http://dx.doi.org/10.21024/pnuedi.28.4.201812.461 * - 2008 2018 - A Study on the Change of Issues with Adolescent Problem

More information

NLTK 6: 텍스트 분류 학습 (` `%%%`#`&12_`__~~~ౡ氀猀攀)

NLTK 6: 텍스트 분류 학습 (` `%%%`#`&12_`__~~~ౡ氀猀攀) nltk.org/book/ch06.html gender_features word last_letter word[-1] word >>> def gender_features(word):... return { last_letter : word[-1]} >>> gender_features( Shrek ) { last_letter : k } nltk.corpus.names

More information

소프트웨어개발방법론

소프트웨어개발방법론 사용사례 (Use Case) Objectives 2 소개? (story) vs. 3 UC 와 UP 산출물과의관계 Sample UP Artifact Relationships Domain Model Business Modeling date... Sale 1 1..* Sales... LineItem... quantity Use-Case Model objects,

More information

untitled

untitled 전방향카메라와자율이동로봇 2006. 12. 7. 특허청전기전자심사본부유비쿼터스심사팀 장기정 전방향카메라와자율이동로봇 1 Omnidirectional Cameras 전방향카메라와자율이동로봇 2 With Fisheye Lens 전방향카메라와자율이동로봇 3 With Multiple Cameras 전방향카메라와자율이동로봇 4 With Mirrors 전방향카메라와자율이동로봇

More information

Recommender Systems - Beyond Collaborative Filtering

Recommender Systems - Beyond Collaborative Filtering Recommender Systems Beyond Collaborative Filtering Sungjoo Ha May 17th, 2016 Sungjoo Ha 1 / 19 Recommender Systems Problem 사용자가얼마나특정아이템을좋아할지예측해보자. 과거행동을바탕으로 다른사용자와의관계를바탕으로 아이템사이의관계로부터 문맥을살펴보고... Sungjoo

More information

cha4_ocw.hwp

cha4_ocw.hwp 제 4장 확률 우리는 일상생활에서 확률이라는 용어를 많이 접하게 된다. 확률(probability)는 한자어로 확실할 확( 確 ), 비율 률( 率 )로 해석된다. 로또당첨확률, 야구 한국시리즈에서 특정 팀이 우승 할 확률, 흡연자가 폐암에 걸릴 확률, 집값이 오를 확률 등 수없이 많은 확률들이 현대생활 에서 사용되어지고 있다. 대부분의 일간신문에는 기상예보

More information

Artificial Intelligence: Assignment 6 Seung-Hoon Na December 15, Sarsa와 Q-learning Windy Gridworld Windy Gridworld의 원문은 다음 Sutton 교재의 연습문제

Artificial Intelligence: Assignment 6 Seung-Hoon Na December 15, Sarsa와 Q-learning Windy Gridworld Windy Gridworld의 원문은 다음 Sutton 교재의 연습문제 Artificial Intelligence: Assignment 6 Seung-Hoon Na December 15, 2018 1 1.1 Sarsa와 Q-learning Windy Gridworld Windy Gridworld의 원문은 다음 Sutton 교재의 연습문제 6.5에서 찾아볼 수 있다. http://incompleteideas.net/book/bookdraft2017nov5.pdf

More information

2 : CNN (Jaeyoung Kim et al.: Experimental Comparison of CNN-based Steganalysis Methods with Structural Differences) (Regular Paper) 24 2, (JBE

2 : CNN (Jaeyoung Kim et al.: Experimental Comparison of CNN-based Steganalysis Methods with Structural Differences) (Regular Paper) 24 2, (JBE 2: CNN (Jaeyoung Kim et al.: Experimental Comparison of CNN-based Steganalysis Methods with Structural Differences) (Regular Paper) 24 2, 2019 3 (JBE Vol. 24, No. 2, March 2019) https://doi.org/10.5909/jbe.2019.24.2.315

More information

Vol.259 C O N T E N T S M O N T H L Y P U B L I C F I N A N C E F O R U M

Vol.259 C O N T E N T S M O N T H L Y P U B L I C F I N A N C E F O R U M 2018.01 Vol.259 C O N T E N T S 02 06 28 61 69 99 104 120 M O N T H L Y P U B L I C F I N A N C E F O R U M 2 2018.1 3 4 2018.1 1) 2) 6 2018.1 3) 4) 7 5) 6) 7) 8) 8 2018.1 9 10 2018.1 11 2003.08 2005.08

More information

2 : (Rahoon Kang et al.: Image Filtering Method for an Effective Inverse Tone-mapping) (Special Paper) 24 2, (JBE Vol. 24, No. 2, March 2019) h

2 : (Rahoon Kang et al.: Image Filtering Method for an Effective Inverse Tone-mapping) (Special Paper) 24 2, (JBE Vol. 24, No. 2, March 2019) h 2: (Rahoon Kang : Image Filtering Method for an Effective Inverse Tone-mapping) (Special Paper) 24 2, 2019 3 (JBE Vol. 24, No. 2, March 2019) https://doi.org/10.5909/jbe.2019.24.2.217 ISSN 2287-9137 (Online)

More information

<32392D342D313020C0FCB0C7BFED2CC0CCC0B1C8F12E687770>

<32392D342D313020C0FCB0C7BFED2CC0CCC0B1C8F12E687770> Journal of the Society of Korea Industrial and Systems Engineering Vol 9 No 4 pp75 8 December 006 유전자 알고리즘을 이용한 시간제약 차량경로문제 * ** * ** 1 Vehicle Routing Problems with Time Window Constraints by Using Genetic

More information

다중 곡면 검출 및 추적을 이용한 증강현실 책

다중 곡면 검출 및 추적을 이용한 증강현실 책 1 딥러닝기반성별및연령대 추정을통한맞춤형광고솔루션 20101588 조준희 20131461 신혜인 2 개요 연구배경 맞춤형광고의필요성 성별및연령별주요관심사에적합한광고의필요성증가 제한된환경에서개인정보획득의한계 맞춤형광고의어려움 영상정보기반개인정보추정 연구목표 딥러닝기반사용자맞춤형광고솔루션구현 얼굴영상을이용한성별및연령대추정 성별및연령대를통합네트워크로학습하여추정정확도향상

More information

example code are examined in this stage The low pressure pressurizer reactor trip module of the Plant Protection System was programmed as subject for

example code are examined in this stage The low pressure pressurizer reactor trip module of the Plant Protection System was programmed as subject for 2003 Development of the Software Generation Method using Model Driven Software Engineering Tool,,,,, Hoon-Seon Chang, Jae-Cheon Jung, Jae-Hack Kim Hee-Hwan Han, Do-Yeon Kim, Young-Woo Chang Wang Sik, Moon

More information

1-1-basic-43p

1-1-basic-43p A Basic Introduction to Artificial Neural Network (ANN) 도대체인공신경망이란무엇인가? INDEX. Introduction to Artificial neural networks 2. Perceptron 3. Backpropagation Neural Network 4. Hopfield memory 5. Self Organizing

More information

歯15-ROMPLD.PDF

歯15-ROMPLD.PDF MSI & PLD MSI (Medium Scale Integrate Circuit) gate adder, subtractor, comparator, decoder, encoder, multiplexer, demultiplexer, ROM, PLA PLD (programmable logic device) fuse( ) array IC AND OR array sum

More information

Naver.NLP.Workshop.SRL.Sogang_Alzzam

Naver.NLP.Workshop.SRL.Sogang_Alzzam : Natra Langage Processing Lab 한국어 ELMo 모델을이용한의미역결정 박찬민, 박영준 Sogang_Azzam Naver NLP Chaenge 서강대학교자연어처리연구실 목차 서론 제안모델 실험 결론 2 서론 의미역결정이란? 문장의술어를찾고, 그술어와연관된논항들사이의의미관계를결정하는문제 논항 : 의미역이부여된각명사구의미역 : 술어에대한명사구의의미역할

More information

목차 AI Boom Chatbot Deep Learning Company.AI s Approach AI Chatbot In Financial service 2

목차 AI Boom Chatbot Deep Learning Company.AI s Approach AI Chatbot In Financial service 2 챗봇과 금융서비스의 결합 2017.05.25 Company.AI 강지훈 목차 1. 2. 3. 4. 5. AI Boom Chatbot Deep Learning Company.AI s Approach AI Chatbot In Financial service 2 3 인공지능 및 고급 기계 학습 딥러닝, 인공신경망, 자연어 처리 등 다양한 기술 이해, 학습, 예측

More information

Microsoft PowerPoint - 알고리즘_5주차_1차시.pptx

Microsoft PowerPoint - 알고리즘_5주차_1차시.pptx Basic Idea of External Sorting run 1 run 2 run 3 run 4 run 5 run 6 750 records 750 records 750 records 750 records 750 records 750 records run 1 run 2 run 3 1500 records 1500 records 1500 records run 1

More information

6주차.key

6주차.key 6, Process concept A program in execution Program code PCB (process control block) Program counter, registers, etc. Stack Heap Data section => global variable Process in memory Process state New Running

More information

Pattern Recognition

Pattern Recognition 딥러닝이해및미디어응용 아주대학교구형일 인공지능 / 기계학습 / 딥러닝 AI 에관한 4 개의관점 Humanly Rationally Thinking Thinking Humanly Thinking Rationally Acting Acting Humanly Acting Rationally Acting Humanly 사람처럼일하는 / 행동하는기계 인공지능은사람에의해서수행될때지능이필요한일을수행하는기계를만드는기술이다.

More information

6 : (Gicheol Kim et al.: Object Tracking Method using Deep Learing and Kalman Filter) (Regular Paper) 24 3, (JBE Vol. 24, No. 3, May 2019) http

6 : (Gicheol Kim et al.: Object Tracking Method using Deep Learing and Kalman Filter) (Regular Paper) 24 3, (JBE Vol. 24, No. 3, May 2019) http (Regular Paper) 24 3, 2019 5 (JBE Vol. 24, No. 3, May 2019) https://doi.org/10.5909/jbe.2019.24.3.495 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a), a), a), b), b), b), a) Object Tracking Method using

More information

Page 2 of 6 Here are the rules for conjugating Whether (or not) and If when using a Descriptive Verb. The only difference here from Action Verbs is wh

Page 2 of 6 Here are the rules for conjugating Whether (or not) and If when using a Descriptive Verb. The only difference here from Action Verbs is wh Page 1 of 6 Learn Korean Ep. 13: Whether (or not) and If Let s go over how to say Whether and If. An example in English would be I don t know whether he ll be there, or I don t know if he ll be there.

More information

Contents I II Project Overview 상황분석 및 여건진단 II-1 문화재청 정책 및 사업 분석 II-2 II-3 문화재청 정책 커뮤니케이션 분석 문화재청 일반인식 분석 III 조직분석 및 사례 연구 III-1 문화재청 홍보 조직 및 예산 분석 III

Contents I II Project Overview 상황분석 및 여건진단 II-1 문화재청 정책 및 사업 분석 II-2 II-3 문화재청 정책 커뮤니케이션 분석 문화재청 일반인식 분석 III 조직분석 및 사례 연구 III-1 문화재청 홍보 조직 및 예산 분석 III 문화재청 중장기 종합홍보전략 수립 연구용역 결과보고 Proposed by INFOMaster 2011. 08. 10 Contents I II Project Overview 상황분석 및 여건진단 II-1 문화재청 정책 및 사업 분석 II-2 II-3 문화재청 정책 커뮤니케이션 분석 문화재청 일반인식 분석 III 조직분석 및 사례 연구 III-1 문화재청 홍보

More information

Product A4

Product A4 2 APTIV Film Versatility and Performance APTIV Film Versatility and Performance 3 4 APTIV Film Versatility and Performance APTIV Film Versatility and Performance 5 PI Increasing Performance PES PPSU PSU

More information

Probability Overview Naive Bayes Classifier Director of TEAMLAB Sungchul Choi

Probability Overview Naive Bayes Classifier Director of TEAMLAB Sungchul Choi Probability Overview Naive Bayes Classifier Director of TEAMLAB Sungchul Choi 머신러닝의학습방법들 - Gradient descent based learning - Probability theory based learning - Information theory based learning - Distance

More information

Ⅱ. Embedded GPU 모바일 프로세서의 발전방향은 저전력 고성능 컴퓨팅이다. 이 러한 목표를 달성하기 위해서 모바일 프로세서 기술은 멀티코 어 형태로 발전해 가고 있다. 예를 들어 NVIDIA의 최신 응용프 로세서인 Tegra3의 경우 쿼드코어 ARM Corte

Ⅱ. Embedded GPU 모바일 프로세서의 발전방향은 저전력 고성능 컴퓨팅이다. 이 러한 목표를 달성하기 위해서 모바일 프로세서 기술은 멀티코 어 형태로 발전해 가고 있다. 예를 들어 NVIDIA의 최신 응용프 로세서인 Tegra3의 경우 쿼드코어 ARM Corte 스마트폰을 위한 A/V 신호처리기술 편집위원 : 김홍국 (광주과학기술원) 스마트폰에서의 영상처리를 위한 GPU 활용 박인규, 최호열 인하대학교 요 약 본 기고에서는 최근 스마트폰에서 요구되는 다양한 멀티미 디어 어플리케이션을 embedded GPU(Graphics Processing Unit)를 이용하여 고속 병렬처리하기 위한 GPGPU (General- Purpose

More information

(JBE Vol. 21, No. 1, January 2016) (Regular Paper) 21 1, (JBE Vol. 21, No. 1, January 2016) ISSN 228

(JBE Vol. 21, No. 1, January 2016) (Regular Paper) 21 1, (JBE Vol. 21, No. 1, January 2016)   ISSN 228 (JBE Vol. 1, No. 1, January 016) (Regular Paper) 1 1, 016 1 (JBE Vol. 1, No. 1, January 016) http://dx.doi.org/10.5909/jbe.016.1.1.60 ISSN 87-9137 (Online) ISSN 16-7953 (Print) a), a) An Efficient Method

More information

SW_faq2000번역.PDF

SW_faq2000번역.PDF FREUENTLY ASKED UESTIONS ON SPEED2000 Table of Contents EDA signal integrity tool (vias) (via) /, SI, / SPEED2000 SPEED2000 EDA signal integrity tool, ( (via),, / ), EDA, 1,, / 2 FEM, PEEC, MOM, FDTD EM

More information

09권오설_ok.hwp

09권오설_ok.hwp (JBE Vol. 19, No. 5, September 2014) (Regular Paper) 19 5, 2014 9 (JBE Vol. 19, No. 5, September 2014) http://dx.doi.org/10.5909/jbe.2014.19.5.656 ISSN 2287-9137 (Online) ISSN 1226-7953 (Print) a) Reduction

More information

sna-node-ties

sna-node-ties Node Centrality in Social Networks Nov. 2015 Youn-Hee Han http://link.koreatech.ac.kr Importance of Nodes ² Question: which nodes are important among a large number of connected nodes? Centrality analysis

More information

(, sta*s*cal disclosure control) - (Risk) and (U*lity) (Synthe*c Data) 4. 5.

(, sta*s*cal disclosure control) - (Risk) and (U*lity) (Synthe*c Data) 4. 5. 1 (, ), ( ) 2 1. 2. (, sta*s*cal disclosure control) - (Risk) and (U*lity) - - 3. (Synthe*c Data) 4. 5. 3 1. + 4 1. 2.,. 3. K + [ ] 5 ' ', " ", " ". (SNS), '. K KT,, KG (PG), 'CSS'(Credit Scoring System)....,,,.

More information

Reinforcement Learning & AlphaGo

Reinforcement Learning & AlphaGo Gait recognition using a Discriminative Feature Learning Approach for Human identification 딥러닝기술및응용딥러닝을활용한개인연구주제발표 이장우 wkddn1108@kist.re.kr 2018.12.07 Overview 연구배경 관련연구 제안하는방법 Reference 2 I. 연구배경 Reinforcement

More information

(JBE Vol. 7, No. 4, July 0)., [].,,. [4,5,6] [7,8,9]., (bilateral filter, BF) [4,5]. BF., BF,. (joint bilateral filter, JBF) [7,8]. JBF,., BF., JBF,.

(JBE Vol. 7, No. 4, July 0)., [].,,. [4,5,6] [7,8,9]., (bilateral filter, BF) [4,5]. BF., BF,. (joint bilateral filter, JBF) [7,8]. JBF,., BF., JBF,. : 565 (Special Paper) 7 4, 0 7 (JBE Vol. 7, No. 4, July 0) http://dx.doi.org/0.5909/jbe.0.7.4.565 a), b), a) Depth Map Denoising Based on the Common Distance Transform Sung-Yeol Kim a), Manbae Kim b),

More information

Slide 1

Slide 1 영상및자연어처리분야인공지능기술동향및전망 주재걸교수 고려대학교정보대학컴퓨터학과 주재걸교수 1 사진 고려대컴퓨터학과 2 3 4 2 3 연구실구성원 9 박사및석박통합과정 22 석사과정 15+ 학부연구생수행과제한국연구재단정보통신기술진흥센터한국산업기술평가관리원한국전력공사삼성리서치마이크로소프트리서치네이버웹툰 SK 텔레콤엔씨소프트자문기관신한금융투자 LG CNS 네이버 Clova

More information

PowerPoint 프레젠테이션

PowerPoint 프레젠테이션 RecurDyn 의 Co-simulation 와 하드웨어인터페이스적용 2016.11.16 User day 김진수, 서준원 펑션베이솔루션그룹 Index 1. Co-simulation 이란? Interface 방식 Co-simulation 개념 2. RecurDyn 과 Co-simulation 이가능한분야별소프트웨어 Dynamics과 Control 1) RecurDyn

More information

SOSCON-MXNET_1014

SOSCON-MXNET_1014 딥러닝계의블루오션, Apache MXNet 공헌하기 윤석찬, Amazon Web Services 오규삼, 삼성 SDS SAMSUNG OPEN SOURCE CONFERENCE 018 목차 11 1 : 4 2 031 1 1 21 51: 1 L 1 S D A 1 S D N M Deep Learning 101 SAMSUNG OPEN SOURCE CONFERENCE

More information

4 CD Construct Special Model VI 2 nd Order Model VI 2 Note: Hands-on 1, 2 RC 1 RLC mass-spring-damper 2 2 ζ ω n (rad/sec) 2 ( ζ < 1), 1 (ζ = 1), ( ) 1

4 CD Construct Special Model VI 2 nd Order Model VI 2 Note: Hands-on 1, 2 RC 1 RLC mass-spring-damper 2 2 ζ ω n (rad/sec) 2 ( ζ < 1), 1 (ζ = 1), ( ) 1 : LabVIEW Control Design, Simulation, & System Identification LabVIEW Control Design Toolkit, Simulation Module, System Identification Toolkit 2 (RLC Spring-Mass-Damper) Control Design toolkit LabVIEW

More information

PowerPoint 프레젠테이션

PowerPoint 프레젠테이션 [ 인공지능입문랩 ] SEOPT ( Study on the Elements Of Python and Tensorflow ) 인공지능 + 데이터분석목적 / 방법 / 기법 / 도구 + Python Programming 기초 + NumpyArray(Tensor) youngdocseo@gmail.com 1 *3 시간 / 회 구분일자내용비고 1 회 0309

More information

Chapter4.hwp

Chapter4.hwp Ch. 4. Spectral Density & Correlation 4.1 Energy Spectral Density 4.2 Power Spectral Density 4.3 Time-Averaged Noise Representation 4.4 Correlation Functions 4.5 Properties of Correlation Functions 4.6

More information

ePapyrus PDF Document

ePapyrus PDF Document 막힌 부분을 갖는 네트워크 내 효과적인 경로 탐색을 위한 유전 알고리즘 적용 김준우 *, 이민정 ** 요약 자연계의 진화 과정을 모방하는 유전 알고리즘은 다양한 조합 최적화와 같은 NP-hard 문제의 해를 탐색하는데 매 우 유용한 도구이다. 본 논문은 네트워크 내에 존재하는 두 노드 사이의 최단 경로를 구하는 문제 풀이를 위하여 유 전 알고리즘을 적용하고자

More information

untitled

untitled Logic and Computer Design Fundamentals Chapter 4 Combinational Functions and Circuits Functions of a single variable Can be used on inputs to functional blocks to implement other than block s intended

More information

Disclaimer IPO Presentation,. Presentation...,,,,, E.,,., Presentation,., Representative...

Disclaimer IPO Presentation,. Presentation...,,,,, E.,,., Presentation,., Representative... DEXTER STUDIOS INVESTOR RELATIONS 2015 Disclaimer IPO Presentation,. Presentation...,,,,, E.,,., Presentation,., Representative... Contents Prologue 01 VFX 02 China 03 Investment Highlights 04 Growth Engine

More information

ecorp-프로젝트제안서작성실무(양식3)

ecorp-프로젝트제안서작성실무(양식3) (BSC: Balanced ScoreCard) ( ) (Value Chain) (Firm Infrastructure) (Support Activities) (Human Resource Management) (Technology Development) (Primary Activities) (Procurement) (Inbound (Outbound (Marketing

More information

PowerPoint 프레젠테이션

PowerPoint 프레젠테이션 ㆍ Natural Language Understanding 관련기술 ㆍ Semantic Parsing Conversational AI Natural Language Understanding / Machine Learning ㆍEntity Extraction and Resolution - Machine Learning 관련기술연구개발경험보유자ㆍStatistical

More information

/ TV 80 () DAB 2001 2002 2003 2004 2005 2010 Analog/Digital CATV Services EPG TV ( 60 ) TV ( Basic, Tier, Premiums 60 ) VOD Services Movies In Demand ( 20 ) Education N- VOD (24 ) Digital Music

More information

- 이 문서는 삼성전자의 기술 자산으로 승인자만이 사용할 수 있습니다 Part Picture Description 5. R emove the memory by pushing the fixed-tap out and Remove the WLAN Antenna. 6. INS

- 이 문서는 삼성전자의 기술 자산으로 승인자만이 사용할 수 있습니다 Part Picture Description 5. R emove the memory by pushing the fixed-tap out and Remove the WLAN Antenna. 6. INS [Caution] Attention to red sentence 3-1. Disassembly and Reassembly R520/ 1 2 1 1. As shown in picture, adhere Knob to the end closely into the arrow direction(1), then push the battery up (2). 2. Picture

More information

Pattern Recognition

Pattern Recognition SELF-DRIVING CARS AND DEEP LEARNING 아주대학교구형일 Course overview (keywords) Introduction Self Driving Cars/Machine Learning/Deep Learning Machine Learning Artificial Neural Network (ANN,MLP) Convolution Neural

More information