Transcription of 画像診断とAI(人工知能)
1 Artificial Intelligence AI Apple Siri Google AI John McCarthy It is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable 1 AI 2016 3 Google DeepMind alphago 4 1 AI 2 alphago Deep Learning machine learning ML 3 AI ML AI AI natural language processing.
2 NLP bioinformatics CT MRI PET 4 1 1980 Computer aided/assisted detection/diagnosis CAD CADe CADx 2000 41 3 19 2019 AI Radiology and Artificial IntelligenceSeiichi KawamotoKey words CAD , Artificial Intelligence, Deep Learning, Computer aided/assisted detection/diagnosis CAD , Radiology, Radiomics, Radiogenomics4 5 5.
3 6 CAD feature representation learning 50 7 AI AI Geoffrey Hinton 2016 Machine Learning and the Market for Intelligence conference 5 10 AI 8 AI AI AI 1960 2 ML Supervised learning Unsupervised learning Reinforcement Learning 3 1 1
4 Latent variable 9 Semi-supervised Learning 10 labeled data unlabeled data [1] 11 12,13 reinforcement learning alphago 14 [1] 5 9 Regression [2] [3]
5 2 Support Vector Machine SVM generalization 2 SVM 15 decision tree 1 2 classification tree regression tree Gradient Tree Boosting Gradient Boosting Random Forest [4] Na ve Bayes Nearest neighbor NN clustering internal cohesion external isolation k k means exploratory dimensionality reduction [2]
6 [3] [4] weak learner= a b 6 Internal representation[5] 1 Artificial neuron 2 1957 Rosenblatt 16 feedforward neural network Multilayer neural network 1 [6]
7 Back Propagation 1986 Rumelhart 17 3 Feedforward Neural network FFNN 1 NN NN 2 NN NN DNN 2010 Feature engineering ML feature Internal representation ML Feature engineering feature feature ML Over training over [5] [6]
8 7fitting DropOut 19 Deep Learning Deep Neural Network DNN 20 Hinton 20 2006 Restricted Boltzmann Machine RBM local optimum [7] Internal representation feature ICT GPU Graphic Processing Unit GPGPU General-purpose computing on Graphics Processing Units HPC High-performance computing 2010 21 Convolutional
9 Neural network CNN Recurrent Neural network RNN ReLU MaxOut DropOut Adam Convolutional neural network CNN 2012 ILSVRC ImageNet Large Scale Visual Recognition Challenge Hinton 22 SuperVision CNN [7] (a) ReLU x 0, .. x 1 Relu .(b),(c) DropOut [18]
10 8 CNN CNN fully connected layer deep ReLU convolution layer feature map 1 RGB3 3 CNN pooling layer CNN max pooling FC CNN