Schedule and Syllabus

본 페이지는 2017년 여름 방학 동안 CNN 공부를 위한 HCIL 내부 세미나 일정입니다.
Stanford University의 CS231n: Convolutional Neural Networks for Visual Recognition (Winter 2016)에 기반하였습니다.
Assignment의 경우, 부가적으로 Spring 2017 강의를 참고하시면 도움이 될 수 있습니다.

1. [동영상 강의 URL]
2. [과제 제출 및 취합 URL]
3. [발표 스케줄 URL]

세미나 기간: 6월 19일 ~ 8월 31일
세미나 시간: 매주 목요일 오후 1시
과제 제출 시간: Due date의 오후 6시 (시간 엄수)
Event TypeDue DateDescriptionCourse Materials
Lecture 1 June 28 Intro to Computer Vision, historical context. [slides]
Lecture 2 June 28 Image classification and the data-driven approach
k-nearest neighbor
Linear classification I		 [slides] [video]
[python/numpy tutorial]
[image classification notes]
[linear classification notes]
A1-part1 Due June 28 Assignment #1 (Q1, Q2) Due date [Assignment #1]
Lecture 3 July 5 Linear classification II
Higher-level representations, image features
Optimization, stochastic gradient descent		 [slides] [video]
[linear classification notes]
[optimization notes]
Lecture 4 July 5 Backpropagation
Introduction to neural networks		 [slides] [video]
[backprop notes]
[Efficient BackProp] (optional)
related: [1], [2], [3] (optional)
A1-part2 Due July 5 Assignment #1 (Q3, Q4, Q5) Due date [Assignment #1]
Lecture 5 July 12 Training Neural Networks Part 1
activation functions, weight initialization, gradient flow, batch normalization
babysitting the learning process, hyperparameter optimization 		[slides] [video]
Neural Nets notes 1
Neural Nets notes 2
Neural Nets notes 3
tips/tricks: [1], [2], [3] (optional)
Deep Learning [Nature] (optional)
Lecture 12 July 12 Overview of Caffe/Torch/Theano/TensorFlow
TensorFlow, PyTorch 중 1개 이상을 선택적으로 학습 (2016, 2017 slide 모두 참고). 		[slides] [2017 slides]
A2-part1 Due July 12 Assignment #2 (Q1) Due date [Assignment #2]
Lecture 6 July 26 Training Neural Networks Part 2: parameter updates, ensembles, dropout
Convolutional Neural Networks: intro 		[slides] [video]
Neural Nets notes 3
Lecture 7 July 26 Convolutional Neural Networks: architectures, convolution / pooling layers
Case study of ImageNet challenge winning ConvNets 		[slides] [video]
ConvNet notes
A2-part2 Due July 26 Assignment #2 (Q2, Q3) Due date [Assignment #2]
Lecture 8 Aug 2 ConvNets for spatial localization
Object detection		 [slides] [video]
Lecture 9 Aug 2 Understanding and visualizing Convolutional Neural Networks
Backprop into image: Visualizations, deep dream, artistic style transfer
Adversarial fooling examples
 [slides] [video]
A2-part3 Due Aug 2 Assignment #2 (Q4, Q5) Due date [Assignment #2]
Lecture 10 Aug 16 Recurrent Neural Networks (RNN), Long Short Term Memory (LSTM)
RNN language models
Image captioning 		[slides] [video]
DL book RNN chapter (optional)
min-char-rnn, char-rnn, neuraltalk2
A3-part1a Due Aug 16 Assignment #3 (Q1-Temporal Affine layer 전까지) Due date [Assignment #3]
Lecture 11 Aug 23 Training ConvNets in practice
Data augmentation, transfer learning
Distributed training, CPU/GPU bottlenecks
Efficient convolutions
 [slides] [video]
A3-part1b Due Aug 23 Assignment #3 (Q1-Temporal Affine layer 부터 끝까지) Due date [Assignment #3]
Proposal Aug 30 이후 본인 연구에 어떻게 적용할지를 Slide에 1페이지로 정리 [제출]
Lecture 13 TBD Segmentation
Soft attention models
Spatial transformer networks
 [slides] [video]
A3-part2 Due TBD Assignment #3 (Q2) Due date [Assignment #3]
Lecture 14 TBD ConvNets for videos
Unsupervised learning
 [slides] [video]
A3-part3-5 Due TBD Assignment #3 (Q3~Q5) Due date [Assignment #3]
Lecture 15 TBD Invited Talk: Jeff Dean [video] Optional