Crowdly
Add to Chrome
Data Analytics (Eng) / Data Analitika (Ing) - 344
Looking for Data Analytics (Eng) / Data Analitika (Ing) - 344 test answers and solutions? Browse our comprehensive collection of verified answers for Data Analytics (Eng) / Data Analitika (Ing) - 344 at stemlearn.sun.ac.za.
Get instant access to accurate answers and detailed explanations for your course questions. Our community-driven platform helps students succeed!
Which of the following is a valid net function output that is acceptable by the tanh activation function? Select all that apply. Negative marking will be used for incorrect answers.
0%
0%
0%
0%
View this question
A fully connected neural network is presented with a batch of 10 instances. The network has 3 sensing neurons, 5 hidden processing neurons and 2 output neurons. How many output values will be in the network output matrix for this batch?
View this question
A fully connected neural network comprising of 8 neurons (neuron1 to neuron8) is presented with a batch of 10 instances. The first 3 neurons in the network are sensing neurons (neuron1, neuron2 and neuron3). The next 3 neurons in the network are hidden layer processing neurons (neuron4, neuron5 and neuron6). The last 2 neurons in the network are output neurons (neuron7 and neuron8). The notation for representing the weights is w(n(l),n(l-1)), where l is the layer number, n(l) represents the number of a node in layer l and n(l-1) represents the number of a node or bias in layer l-1. Which of the following weights do not exist in this network? Select all that apply. Negative marking will be used for incorrect answers.
0%
0%
0%
0%
View this question
A fully connected neural network is presented with a batch of 10 instances. The network has 2 sensing neurons, 5 hidden processing neurons and 1 output neuron. How many weights does this network have?
View this question
Which of the following activation functions is non-linear and has a differential whose value can be greater than 1? Select all that apply. Negative marking will be used for incorrect answers.
0%
100%
0%
0%
View this question
Which of the following represent a valid value of a bias weight in a neural network? Select all that apply. Negative marking will be used for incorrect answers.
100%
100%
100%
100%
View this question
True or false? The depth of a deep neural network is a parameter since depth influences the predictive performance of the network.
0%
0%
View this question
True or false? f(x) = |x| represents a linear function for values of x in the range -10 to -1.
0%
0%
View this question
A neural network model predicts an output vector of (02, 0.5, 0.3) for an instance whose one-hot-encoded target vector is (0,1,0). What is the cross entropy loss for this prediction, correct to 2 decimal places? Assume the cross entropy uses the natural logarithm.
View this question
True or false? The point in the feature space, at which stochastic gradient descent reaches the global minimum of a neural network's error surface, corresponds with an L2 error of 0 for the test instance being fed through the network.
0%
100%
View this question