Deep Learning Quiz 2: 2 April 2023, Set QPE2 (January 2023 term)

1. A

   AdaGrad takes lesser number of iterations than vanilla Gradient Descent toconverge

2. B

   AdaGrad takes more number of iterations than vanilla Gradient Descent toconverge

3. C

   Both the algorithms will converge in equal number of iterations

4. D

   Both the algorithms will keep oscillating around the minimum

1. A

   True, due to the presence of noise in the target

2. B

   True, due to the high variance of the prediction model

3. C

   False, due to the high variance of the prediction model

4. D

   False, due to zero mean of the noise added to the target

Select all that apply.

1. A

   They are more suitable for sparse data

2. B

   They are more suitable for dense data

3. C

   They add additional computational complexity to the vanilla gradient descentalgorithm

4. D

   They are guranteed to converge to global minimum