Neural Network and Gradient Descent Quiz - 2

Why convolution neural network is taking off quickly in recent times?

• Availability of more computational power

• Integration of feature extraction within the training process

• Access to large amount of digitized data

• All of the above

Momentum based gradient descent algorithm and Nesterov accelerated gradient descent are faster than Stochastic gradient descent algorithm?

• Not mentioned

• True

• False

• Not applicable

Consider the following statement, 'It takes less time to navigate the regions having a gentle slope'. The above statement is true in case of

• Not mentioned

• Both I and II

• Momentum based gradient descent algorithm

• Gradient descent algorithm

Identify the technique that is used to achieve relatively better learning rate by updating w using bunch of different values of η.

• All the above

• Line Search

• Bias Correction

• Stochastic

There is no guarantee that the loss decreases at each step in a stochastic Gradient Descent?

• False

• True

Identify the advantages of Nesterov accelerated gradient descent.

• Chances of escaping minima valley are also smaller

• Corrects its course quicker than Momentum-based gradient descent

• Oscillations are smaller

• All of the above

Pick out the methods for annealing learning rate that has only number of epochs as the hyperparameter.

• Step decay

• Not mentioned

• Exponential Decay

• 1/t Decay

Adagrad got stuck when it was close to convergence. How does RMSProp overcome this problem?

• More Aggressive on decay

• Less Aggressive on decay

• No decay

Which of the following gradient descent algorithm suffers from more oscillations?

• Momentum based gradient descent

• Nesterov accelerated gradient descent

• Vanilla gradient descent

• None of the above

In a neural network, knowing the weight and bias of each neuron is the most important step. If you

can somehow get the correct value of weight and bias for each neuron, you can approximate any

function. What would be the best way to approach this?

• Assign random values and pray to God they are correct

• Search every possible combination of weights and biases till you get the best value

• Iteratively check that after assigning a value how far you are from the best values, and slightly change the assigned values values to make them better

• None of these