("Note: This is an interactive item Follow the instructions to answer the question Scroll bars may appear it the windows are too smart Each window can be resized by dragging on the 5 circles located between windows)
You are building a CNN for an image classification task
Drag the layers on the left to the slots on the right in the appropriate order (from top to bottom) for this task
You have the number of layers listed below:
Input Layer (1)
Output Layer (1)
Convolutional Layer (1)
Fully Connected Layer(2>
Pooling Layer(1)
Select and Place:
Given these two addLayer action calls:
How many trainable parameters for layer fc2?
A. 70
B. 1050
C. 1020
D. 1000
Which option is the correct activation (unction for the output layer in a CNN model trained to classify an image belonging to one of the n classes (CI. C2, C3, , Cn)?
A. Sigmoid
B. ReLU
C. TanH
D. Softmax
Which statement is TRUE concerning the "dropout" option?
A. It specifies the number of neurons to drop from each layer of the network
B. It specifies the percentage of neurons to drop from a given layer of the network
C. It specifies the percentage of neurons to drop from the entire network
D. It specifies the number of layers to drop from the entire network
What are two reasons to use a pooling layer in a Convolution Neural Network (CNN)? (Choose two)
A. to increase granularity of information
B. to make the feature map invariant to transformation
C. to extract features of different spatial scale
D. to extract a summarized value from an area of the image
Given the code:
How many trainable parameters do these lines add to the model
A. 25
B. 520
C. 250
D. 500
Consider the CONCEPT rule:
CONCEPT:red@A
Which choice represents a possible result from this CONCEPT rule?
A. Redraw begins with the word red and then matches any set of alpha characters
B. Reds, a plural variant of the noun red
C. Redness a noun having red as a root word
D. Redder, a variant of the adjective red
Complete the concept rule by adding the Boolean Operator that should be used to find concepts where the term "best" appears within five tokens of the term "picture"
Enter your answer in the field above.
A. DIST_5
What computation is performed when applying a neural network with Batch Normalization to test data?
A. μ and o^2 are estimated using an exponentially weighted average across tram and test sets
B. Normalizations are applied using parameters learned during testing
C. Normalizations are applied using parameters learned during training
D. μ and o^2 are estimated using an exponentially weighted average across test set
Which statement describes a situation in which a bi-directional recurrent neural network is a better option than a unidirectional network?
A. The data at any given point in time depends on both the previous and subsequent elements of the sequence
B. The data at any given point in time depends only on the previous elements of the sequence
C. The data at any given point in time is independent of both the previous and subsequent elements of the sequence
D. The data at any given point in time depends only on the subsequent elements of the sequence
Refer to the exhibit.
What is the false negative rate for the category that has the 2nd worst average measure1?
A. Between 1 and 2 B. Less than 1
C. Between 2 and 3
D. Greater than 5
What sampling method does the DLTUNE action use to sample the hyper parameter values?
A. Simple Random Sampling
B. Systematic Sampling
C. Stratified Random Sampling
D. Latin Hypercube Sampling
Which statements are TRUE regarding SAS Visual Text Analytics projects? (Choose two)
A. Users can include nodes from other Model Studio projects, including SAS Visual Forecasting and SAS Visual Data Mining and Machine B. Learning projects, in a SAS Visual Text Analytics pipeline
B. After creating a project in SAS Visual Text Analytics users can run the project with a different data source
C. Users can promote a SAS Visual Text Analytics 8 2 project to SAS Visual Text Analytics 8 3 project
D. A SAS Visual Text Analytics project created by one user cannot be accessed by another user
Which activation function should be used to obtain a single scalar with value in (-1,1)9
A. TanH
B. Logistic
C. Softmax
D. ReLU
Refer to the exhibit.
The pixel intensity values for the top left corner of a 224x224 grey scale image are provided in Exhibit A This image is provided as input to the convolutional filter (Exhibit B) with equal zero padding of size 2 on all sides
What is the first value in the feature map resulting from applying this filler (cross-correlation operation)?
A. 186
B. 0
C. 10
D. 97