Convolutional Neural Networks

Five questions. Choose one answer for each, then submit to see your score and explanations.

The practice target is 80%. You can retry as often as you like, and every lesson stays accessible.

You can practice as a guest. Sign in before starting an attempt to save your scores.

1. What does weight sharing mean in a convolutional layer?A. Every pixel gets its own independent networkB. The training and test sets share labelsC. The same filter weights are applied at different spatial locationsD. Every layer must use the same filter

2. What does flattening a feature tensor do?A. Deletes spatial positions with negative valuesB. Rearranges its values into a vector without averaging themC. Keeps only the maximum from each regionD. Learns a new filter automatically

3. How does max pooling differ from average pooling?A. It selects a regional maximum instead of computing a regional meanB. It always increases the image dimensionsC. It learns the mean as a trainable weightD. It replaces all values with class probabilities

4. A 5-by-5 input uses a 3-by-3 filter, stride 1, and no padding or dilation. What is the output size?A. 5 by 5B. 2 by 2C. 7 by 7D. 3 by 3

5. A 2-by-2 region contains 1, 5, 2, and 4. What value does 2-by-2 max pooling produce?A. 12B. 1C. 5D. 3

Review this module from its first lesson →

Convolutional Neural Networks

Your result