In stochastic gradient descent the high variance frequent parameter updates causes the loss function to fluctuate heavily.

Stochastic gradient descent is also known as on-line gradient descent.

A

True

B

False

Which of the following statements is not true about stochastic gradient descent for regularised loss minimisation?

A

Stochastic gradient descent has the same worst-case sample complexity bound as regularised loss minimisation

B

On some distributions, regularised loss minimisation yields a better solution than stochastic gradient descent

C

In some cases we solve the optimisation problem as associated with regularised loss minimisation

D

Stochastic gradient descent has entirely different worst-case sample complexity bound from regularised loss minimisation

In each question below is given a statement numbered I, II and III. An assumption is something supposed or taken for granted. You have to consider the following assumption and decide which of the assumption is implicit in the statement. Statement: Every time Sri Lanka’s president, Mahinda Rajapaksa, turned to his Chinese allies for loans and assistance with an ambitious port project, the answer was yes. Yes, though feasibility studies said the port wouldn’t work. Yes, though other frequent lenders like India had refused. Yes, though Sri Lanka’s debt was ballooning rapidly under Rajapaksa. Assumptions: I. China had a stake in Sri Lanka’s main port as well, China Harbor was building a new terminal there, known at the time as Colombo Port City. Along with that deal came roughly 50 acres, solely held by the Chinese company, where Sri Lanka had no sovereignty. II. Sri Lankan officials are quick to point out that the agreement explicitly rules out China’s military use of the site. But others also note that Sri Lanka’s government, still heavily indebted to China, could be pressured to allow it. III. The rising debt and project costs, even as the port was struggling, handed Sri Lanka’s political opposition a powerful issue, and it campaigned heavily on suspicions about China. Rajapaksa may lose the election.

A

Only I is implicit

B

All except I, II is implicit

C

Neither I, II nor III is implicit

D

Only II is implicit

E

Both III and II are implicit

Which of the following statements is true about stochastic gradient descent?

A

It processes one training example per iteration

B

It is not preferred, if the number of training examples is large

C

It processes all the training examples for each iteration of gradient descent

D

It is computationally very expensive, if the number of training examples is large

Stochastic gradient descent (SGD) methods handle redundancy in the data much more efficiently than batch methods.

A

True

B

False

Which of the following statements is not true about the stochastic gradient descent?

A

The parameters are being updated after one iteration

B

It is quite faster than batch gradient descent

C

Stochastic gradient descent is faster than mini batch gradient descent

D

When the number of training examples is large, it can be additional overhead for the system

Stochastic gradient descent falls under Non-convex optimization.

A

True

B

False

Which of the following statements is not true about stochastic gradient descent?

A

Due to the frequent updates, there can be so many noisy steps

B

It may take longer to achieve convergence to the minima of the loss function

C

Frequent updates are computationally expensive

D

It is computationally slower

Stochastic gradient descent has the possibility of escaping from local minima.

A

False

B

True

Stochastic gradient descent cannot be used for risk minimisation.

A

False

B

True

In stochastic gradient descent the high variance frequent parameter updates causes the loss function to fluctuate heavily.

Related Questions