What statements are true?Group of answer choicesBy the end of our network, if we have multiple convolution and pooling layers, our final image is smaller than the original input image and we can classify the entire image, not just a region.Two common pooling methods are max pooling and mean pooling.If you use Photoshop to switch the location of the eyes and the mouth of a cat, a convolutional net with pooling layers would likely still "recognize" the image as a cat.Pooling introduces translation invariance by conveying what patterns are apparent in the image, without conveying where they are exactly.

In a convolutional neural network (CNN), what is the primary role of the pooling layer? Question 2 options: Reducing the number of parameters in the network. Introducing non-linearities into the network. Performing element-wise multiplication with the filter weights. Down-sampling the feature maps to reduce spatial dimensions.

11. Which of the following is a correct statement about a CNN (Convolution Neural Network) structure?以下哪個是關於卷積神經網路架構的正確說法？Number classification and identification is the main function of convolution layers and pooling layers, while fully connected layers are mainly used for image feature extraction. 卷積層與池化層的功能主要是數字分類辨識，而全連結層則主要用於擷取影像的特徵。There have to be more convolution layers than pooling layers. 卷積層的數量需要比池化層多。There have to be more fully connected layers than convolution and pooling layers. 全連接層的數量需要比卷積層與池化層多。Convolution layers are used to extract features from images, while fully connected layers are mainly used to make a classification decision. 卷積層主要是用來擷取影像的特徵，而全連結層則主要用於決定如何分類。

9. If you are now using a Convolution Neural Network (CNN) to distinguish Aaron Kwok and Andy Lau, which of the following is a typical data flow?你現在需要利用卷積神經網路判斷郭富城和劉德華，以下哪個是典型的數據流程呢？Image Input → Convolution Layer → Pooling Layer → Pooling Layer → Fully Connected Layer → Output 影像輸入 → 卷積層 → 池化層 → 池化層 → 全連接層 → 輸出Image Input → Pooling Layer → Convolution Layer → Fully Connected Layer → Output 影像輸入 → 池化層 → 卷積層 → 全連接層 → 輸出Image Input → Convolution Layer → Pooling Layer → Convolution Layer → Pooling Layer → Fully Connected Layer → Output 影像輸入 → 卷積層 → 池化層 → 卷積層 → 池化層 → 全連接層 → 輸出Image Input → Convolution Layer → Pooling Layer → Fully Connected Layer → Convolution Layer → Pooling Layer → Fully Connected Layer → Output 影像輸入 → 卷積層 → 池化層 → 全連接層 → 卷積層 → 池化層 → 全連接層 → 輸出

Which statement is true?Group of answer choicesNeural nets are not great for image data because for images, you only need to learn a few parametres but neural nets can learn many.Small changes in images don’t cause massively different image vectors.Images are not very high dimension and thus, are not very painful to train.CNNs are great for images because they are translation invariant.

Question 10Why does transfer learning work?1 pointAll layers of filters can be learned by studying the mammalian receptive fields.All images are composed of pixels with three color channels.Low-level features are specialized for a particular task, while top-level features are universal to all images.Top-level features are specialized for a particular task, while low-level features are universal to all images