Permalink
Find file
Fetching contributors…
Cannot retrieve contributors at this time
253 lines (219 sloc) 7.9 KB
---
layout: default
---
<div>
스탠포드 CS231n 강의 <a href="http://cs231n.stanford.edu/">CS231n: Convolutional Neural Networks for Visual Recognition</a>에 대한 강의노트의 한글 번역 프로젝트입니다.
<br>
질문/논의거리/이슈 등은 <a href="mailto:team.aikorea@gmail.com">AI Korea 이메일</a>로 연락주시거나, <a href="https://github.com/aikorea/cs231n.github.io">GitHub 레포지토리</a>에 pull request, 또는 이슈를 열어주세요.
</div>
<div class="home">
<div class="materials-wrap">
<div class="module-header">
<a href="glossary/">Glossary</a>
</div>
<div class="module-header">
<a href="video-lectures/">강의 동영상</a>
</div>
<div class="module-header">Winter 2016 과제</div>
<div class="materials-item">
<a href="assignments2016/assignment1/">
Assignment #1: 이미지 분류, kNN, SVM, Softmax, 뉴럴 네트워크
</a>
<span style="float:right">
<progress value="4" max="6"></progress>
</span>
</div>
<div class="materials-item">
<a href="assignments2016/assignment2/">
Assignment #2: Fully-Connected 네트워크, 배치 정규화(Batch Normalization), Dropout,
컨볼루션 신경망
</a>
<span style="float:right">
<progress value="1" max="5"></progress>
</span>
</div>
<div class="materials-item">
<a href="assignments2016/assignment3/">
Assignment #3: 회귀신경망(Recurrent Neural Networks), 이미지 캡셔닝(Captioning),
이미지 그라디언트, DeepDream
</a>
<span style="float:right">
<progress value="20" max="100"></progress>
</span>
</div>
<!--
<div class="module-header">Winter 2015 Assignments</div>
<div class="materials-item">
<a href="assignment1/">
Assignment #1: Image Classification, kNN, SVM, Softmax
</a>
</div>
<div class="materials-item">
<a href="assignment2/">
Assignment #2: Neural Networks, ConvNets I
</a>
</div>
<div class="materials-item">
<a href="assignment3/">
Assignment #3: ConvNets II, Transfer Learning, Visualization
</a>
</div>
-->
<div class="module-header">Module 0: 준비</div>
<div class="materials-item">
<a href="python-numpy-tutorial/">
Python / Numpy Tutorial
</a>
<span style="float:right" class="progress">
Complete! <progress value="1099" max="1099"></progress>
</span>
</div>
<div class="materials-item">
<a href="ipython-tutorial/">
IPython Notebook Tutorial
</a>
<span style="float:right;" class="progress">
Complete! <progress value="100" max="100"></progress>
</span>
</div>
<div class="materials-item">
<a href="terminal-tutorial/">
Terminal.com Tutorial
</a>
<span style="float:right" class="progress">
Complete! <progress value="100" max="100"></progress>
</span>
</div>
<div class="materials-item">
<a href="aws-tutorial/">
AWS Tutorial
</a>
<span style="float:right" class="progress">
Complete! <progress value="100" max="100"></progress>
</span>
</div>
<!-- hardcoding items here to force a specific order -->
<div class="module-header">Module 1: 신경망 구조</div>
<div class="materials-item">
<a href="classification/">
이미지 분류: 데이터 기반 방법론, k-Nearest Neighbor, train/val/test 구분
</a>
<span style="float:right" class="progress">
Complete! <progress value="294" max="294"></progress>
</span>
<div class="kw">
L1/L2 거리, hyperparameter 탐색, 교차검증(cross-validation)
</div>
</div>
<div class="materials-item">
<a href="linear-classify/">
선형 분류: Support Vector Machine, Softmax
</a>
<span style="float:right" class="progress">
<progress value="73" max="370"></progress>
</span>
<div class="kw">
parameteric 접근법, bias 트릭, hinge loss, cross-entropy loss, L2 regularization, 웹 데모
</div>
</div>
<div class="materials-item">
<a href="optimization-1/">
최적화: 확률 그라디언트 하강(Stochastic Gradient Descent)
</a>
<span style="float:right" class="progress">
Complete! <progress value="100" max="100"></progress>
</span>
<div class="kw">
'지형'으로서의 최적화 목적 함수 (optimization landscapes), 국소 탐색(local search), 학습 속도(learning rate), 해석적(analytic)/수치적(numerical) 그라디언트
</div>
</div>
<div class="materials-item">
<a href="optimization-2/">
Backpropagation, 직관
</a>
<span style="float:right" class="progress">
<progress value="240" max="300"></progress>
</span>
<div class="kw">
연쇄 법칙 (chain rule) 해석, real-valued circuits, 그라디언트 흐름의 패턴
</div>
</div>
<div class="materials-item">
<a href="neural-networks-1/">
신경망 파트 1: 네트워크 구조 정하기
</a>
<span style="float:right" class="progress">
<progress value="32" max="220"></progress>
</span>
<div class="kw">
생물학적 뉴런 모델, 활성 함수(activation functions), 신경망 구조, 표현력(representational power)
</div>
</div>
<div class="materials-item">
<a href="neural-networks-2-kr/">
신경망 파트 2: 데이터 준비 및 Loss
</a>
<span style="float:right" class="progress">
<progress value="232" max="308"></progress>
</span>
<div class="kw">
전처리, weight 초기값 설정, 배치 정규화(batch normalization), regularization (L2/dropout), 손실함수
</div>
</div>
<div class="materials-item">
<a href="neural-networks-3/">
신경망 파트 3: 학습 및 평가
</a>
<span style="float:right" class="progress">
<progress value="390" max="390"></progress>
</span>
<div class="kw">
그라디언트 체크, 버그 점검, 학습 과정 모니터링, momentum (+nesterov), 2차(2nd-order) 방법, Adagrad/RMSprop, hyperparameter 최적화, 모델 ensemble
</div>
</div>
<!--
<div class="materials-item">
<a href="neural-networks-case-study/">
Putting it together: Minimal Neural Network Case Study
</a>
<div class="kw">
minimal 2D toy data example
</div>
</div>
-->
<div class="module-header">Module 2: Convolutional Neural Networks</div>
<div class="materials-item">
<a href="convolutional-networks/">
컨볼루션 신경망: 구조, Convolution / Pooling 레이어들
</a>
<span style="float:right" class="progress">
Complete! <progress value="100" max="100"></progress>
</span>
<div class="kw">
레이어(층), 공간적 배치, 레이어 패턴, 레이어 사이즈, AlexNet/ZFNet/VGGNet 사례 분석, 계산량에 관한 고려 사항들
</div>
</div>
<div class="materials-item">
<a href="understanding-cnn/">
컨볼루션 신경망 분석 및 시각화
</a>
<span style="float:right" class="progress">
<progress value="14" max="107"></progress>
</span>
<div class="kw">
tSNE embeddings, deconvnets, 데이터에 대한 그라디언트, ConvNet 속이기, 사람과의 비교
</div>
</div>
<div class="materials-item">
<a href="transfer-learning/">
Transfer Learning and Fine-tuning Convolutional Neural Networks
</a>
<span style="float:right" class="progress">
<progress value="0" max="100"></progress>
</span>
</div>
<div class="module-header">
<a href="acknowledgement/">Acknowledgement</a>
</div>
</div>
</div>