What is the goal of the backpropagation algorithm?
A. to randomize the trajectory of the neural network parameters during training
B. to smooth the gradient of the loss function in order to avoid getting trapped in small local minimas
C. to scale the gradient descent step in proportion to the gradient magnitude
D. to compute the gradient of the loss function with respect to the neural network parameters
The formula for recall is given by (True Positives) / (True Positives + False Negatives).
What is the recall for this example?
A. 0.2
B. 0.25
C. 0.5
D. 0.33
What is used to scale large positive values during data cleaning?
A. division by random numbers
B. square
C. logarithm
D. subtract median
Which measure can be used to show business stakeholders the likelihood that a machine learning model will generate a true prediction?
A. accuracy
B. variance
C. mean
D. skewness
Which IBM Watson Machine Learning deployment method offers the ultimate flexibility in deploying a machine learning model?
A. Watson Machine Learning Python client
B. Watson Machine Learning FORTRAN client
C. Watson Studio Project
D. Watson Machine Learning REST API
Which fine-tuning technique does not optimize the hyperparameters of a machine learning model?
A. grid search
B. population based training
C. random search
D. hyperband
Which statement is true for naive Bayes?
A. Naive Bayes can be used for regression.
B. Let p(C1 | x) and p(C2 | x) be the conditional probabilities that x belongs to class C1 and C2 respectively, in a binary model, log p (C1 | x) ?log p(C2 | x)>; 0 results in predicting that x belongs to C2.
C. Naive Bayes is a conditional probability model.
D. Naive Bayes doesn't require any assumptions about the distribution of values associated with each class.
In a hyperparameter search, whether a single model is trained or a lot of models are trained in parallel is largely determined by?
A. The number of hyperparameters you have to tune.
B. The presence of local minima in your neural network.
C. The amount of computational power you can access.
D. Whether you use batch or mini-batch optimization.
Which statement defines p-value?
A. It is the probability of accepting a null hypothesis when the hypothesis is proven true.
B. It is the probability of rejecting a null hypothesis when the hypothesis is proven false.
C. It is the probability of accepting a null hypothesis when the hypothesis is proven false.
D. It is the probability of rejecting a null hypothesis when the hypothesis is proven true.
Which is the most important thing to ensure while collecting data?
A. samples collected are skewed with each other
B. samples collected are all strongly correlated with each other
C. samples collected adequately cover the space of all possible scenarios
D. samples collected focus only on the most common cases
