Question 1

What is the primary objective of a Denoising Autoencoder?

Options :

A : To reconstruct clean data from noisy inputs

B : To add noise to the input for robust learning

C : To compress data

D : To handle missing values

Answer: A

Question 2

Which optimization algorithm is known for its use of momentum to accelerate convergence and improve gradient descent performance in deep learning?

Options :

A : RMSprop Optimizer adapts learning rates based on recent gradients but does not incorporate momentum.

B : Stochastic Gradient Descent (SGD) with Momentum is used to enhance convergence speed by adding a fraction of the previous update to the current gradient.

C : Adagrad Optimizer adjusts learning rates based on historical gradients but does not use momentum.

D : Adam Optimizer uses adaptive learning rates for different parameters and combines momentum with adaptive gradients for improved training.

Answer: D

Question 3

Which challenge is commonly associated with deploying Edge AI solutions?

Options :

A : Limited computational power and resources on edge devices

B : Complex data integration with central cloud systems

C : High bandwidth requirements for edge-to-cloud data transfer

D : Increased latency due to local data processing

Answer: A

Question 4

What is the primary benefit of using prompt engineering in advanced applications of large language models?

Options :

A : Enhancing the interpretability of outputs

B : Simplifying the model’s training process

C : Reducing the model-s computational load

D : Tailoring model responses to specific tasks

Answer: D

Question 5

Which approach helps in mitigating bias during AI model training?

Options :

A : Increasing the model complexity to handle bias

B : Focusing on high accuracy rather than fairness

C : Applying diverse and representative training datasets

D : Reducing the number of features used in the model

Answer: C