내년엔 꼭!
http://www.pycon.kr/2015/programs/list/
2015년 11월 9일 월요일
Convolutional Neural Networks
Convolutional Neural Network = Convolution + Neural Network
It is basically a neural network, but with a convolution layer.
Simply put, feeding the neural network with preprocessed features not the raw image itself.
Convolution layers do this preprocessing.
Useful Resources:
Deepcumen: http://deepcumen.com/2015/04/convolutional-neural-network/#comment-458
CS231N.Stanford: http://cs231n.github.io/convolutional-networks/
It is basically a neural network, but with a convolution layer.
Simply put, feeding the neural network with preprocessed features not the raw image itself.
Convolution layers do this preprocessing.
Useful Resources:
Deepcumen: http://deepcumen.com/2015/04/convolutional-neural-network/#comment-458
CS231N.Stanford: http://cs231n.github.io/convolutional-networks/
2015년 11월 7일 토요일
Statistical Inference
Statistical inference is about validating how good an estimate is.
There are two basic methods.
1. Hypothesis Testing Using Confidence intervals
- Test if null hypo falls within confidence interval
2. Hypothesis Testing Using P-values
- Test if p value is greater than alpha
Although they sound quite different, they are actually very similar, or even essentially the same.
Testing if an estimate falls within 0.95 confidence interval or if its p value is greater than alpha value of 0.05 is essentially the same logic.
There are two basic methods.
1. Hypothesis Testing Using Confidence intervals
- Test if null hypo falls within confidence interval
2. Hypothesis Testing Using P-values
- Test if p value is greater than alpha
Although they sound quite different, they are actually very similar, or even essentially the same.
Testing if an estimate falls within 0.95 confidence interval or if its p value is greater than alpha value of 0.05 is essentially the same logic.
피드 구독하기:
덧글 (Atom)