이승철
포스텍 기계공학과 교수
UNIST 기계항공및원자력공학부 교수
UNIV. OF MICHIGAN DEPT. OF MECHANICAL ENGINEERING 박사 후 연구원
UNIV. OF MICHIGAN Mechanical Engineering 박사
Deep Learning (딥러닝)
POSTECH
이 강좌는 초대된 사람만 등록가능합니다
강좌 소개
최근 인공지능 분야에서 주목받고 있는 딥러닝을 이해하고 구현한다.
This course is designed to exploit and understand Artificial Intelligence, especially Deep Learning. Students are expected to learn theoretical backgrounds and their implementations of algorithms in Python. Starting from a basic machine learning, various kinds of neural networks will be intensively studied. Numerical Python coding is heavily required during lectures and homework assignments.
평가 방법
- ㅇ퀴즈: 20%
- ㅇ과제: 30%
- ㅇ시험 성적: 50%
- - 프로그래밍 시험 (25%)
- - 필기 시험 (25%)
강좌 목차
| 주차 | 내용 | |
|---|---|---|
| 1 | Introduction & Optimization | Introduction |
| Optimization 1 | ||
| Optimization 2 | ||
| Optimization 3 | ||
| Lab | ||
| 2 | Machine Learning | Regression |
| Classification: Perceptron 1 | ||
| Classification: Perceptron 2 | ||
| Classification: Logistic Regression 1 | ||
| Classification: Logistic Regression 2 | ||
| Lab | ||
3 | Machine Learning with Tensorflow | Machine Learning with Tensorflow 1 |
| Machine Learning with Tensorflow 2 | ||
| Regression & Classification with Tensorflow | ||
| Lab | ||
| 4 | Machine Learning Optimization | Stochastic Gradient Descent 1 |
| Stochastic Gradient Descent 2 | ||
| Lab | ||
| Overfitting | ||
| Lab | ||
| 5 | From Perceptron to MLP (ANN) | Artificial Neural Networks 1 |
| Artificial Neural Networks 2 | ||
| Artificial Neural Networks 3 | ||
| Lab | ||
| Artificial Neural Networks Training 1 | ||
| Artificial Neural Networks Training 2 | ||
| Lab | ||
| 6 | ANN advanced | ANN with Tensorflow 1 |
| ANN with Tensorflow 2 | ||
| Lab | ||
| ANN advanced 1 | ||
| ANN advanced 2 | ||
| Lab | ||
| 7 | Autoencoder (AE) & Convolutional Neural Networks (CNN) | Autoencoder 1 |
| Autoencoder 2 | ||
| Lab | ||
| Convolution: 1D | ||
| Convolution: 2D | ||
| Convolution: Kernel 1 | ||
| Convolution: Kernel 2 | ||
| 8 | Convolutional Neural Networks (CNN) & Class Activation Map (CAM) | Convolution: Padding and Stride |
| Convolution: Pooling | ||
| Convolutional Neural Network in Tensorflow | ||
| Lab | ||
| Class Activation Map (CAM) 1 | ||
| Class Activation Map (CAM) 2 | ||
| Lab |
| 9 | Modern CNNs & Transfer Learning | Modern CNNs |
| Lab | ||
| Transfer Learning | ||
| Lab | ||
| 10 | Convolutional Autoencoders (CAE) & Fully Convolutional Networks (FCN) | Convolutional Autoencoders (CAE) 1 |
| Convolutional Autoencoders (CAE) 2 | ||
| Lab | ||
| Fully Convolutional Networks (FCN) | ||
| Lab | ||
| 11 | Generative Adversarial Networks (GAN) | Generative Adversarial Networks (GAN) 1 |
| Generative Adversarial Networks (GAN) 2 | ||
| Generative Adversarial Networks (GAN) 3 | ||
| Generative Adversarial Networks (GAN) 4 | ||
| 12 | Conditional GAN | Conditional GAN |
| Lab | ||
| 13 | Time Series Analysis | Time Series Data 1 |
| Time Series Data 2 | ||
| Markov Chain | ||
| Hidden Markov Model (HMM) and Kalman Filter | ||
| Lab | ||
| 14 | Recurrent Neural Networks (RNN) | Recurrent Neural Networks (RNN) 1 |
| Recurrent Neural Networks (RNN) 2 | ||
| Recurrent Neural Networks (RNN) 3 | ||
| Recurrent Neural Networks (RNN) 4 | ||
| Recurrent Neural Networks (RNN) 5 | ||
| Recurrent Neural Networks (RNN) 6 | ||
| Recurrent Neural Networks (RNN) 7 | ||
| Lab | ||
| 15 | Final exams | Final exams |
교수자