Deep Learning – Chorale Prelude Interview Question-Answer

Q.1 __________ is a Neural Nets way of classifying inputs.

       A. Classification

       B. Learning

       C. Forward Propagation

       D. Activation

Ans : Learning

Q.2 Recurrent Neural Networks are best suited for Text Processing.

       A. True

       B. False

Ans : True

Q.3 Neural Networks Algorithms are inspired from the structure and functioning of the Human Biological Neuron.

       A. True

       B. False

Ans : True

Q.4 The rate at which cost changes with respect to weight or bias is called _________.

       A. Loss

       B. Gradient

       C. Rate of Change

       D. Derivative

Ans : Gradient

Q.5 __________ is a recommended Model for Pattern Recognition in Unlabeled Data.

       A. Autoencoders

       B. CNN

       C. RNN

       D. Shallow Neural Networks

Ans : Autoencoders

Q.6 What does LSTM stand for?

       A. Least Square Time Mean

       B. Least Squares Term Memory

       C. Long Short Term Memory

       D. Long Short Threshold Memory

Ans : Long Short Term Memory

Q.7 RELU stands for __________.

       A. Rectified Lagrangian Unit

       B. Rectified Linear Unit

       C. Regressive Lagrangian Unit

       D. Regressive Linear Unit

Ans : Rectified Linear Unit

Q.8 All the Visible Layers in a Restricted Boltzmannn Machine are connected to each other.

       A. True

       B. False

Ans : False

Q.9 A Deep Belief Network is a stack of Restricted Boltzmann Machines.

       A. True

       B. False

Ans : True

Q.10 All the neurons in a convolution layer have different Weights and Biases.

       A. True

       B. False

Ans : False

Q.11 A Shallow Neural Network has only one hidden layer between Input and Output layers.

       A. True

       B. False

Ans : True

Q.12 GPU stands for __________.

       A. General Processing Unit

       B. Graphics Processing Unit

       C. Gradient Processing Unit

       D. Good Processing Unit

Ans : Graphics Processing Unit

Q.13 Autoencoders cannot be used for Dimensionality Reduction.

       A. True

       B. False

Ans : True

Q.14 A _______________ matches or surpasses the output of an individual neuron to a visual stimuli.

       A. Convolution

       B. Max Pooling

       C. Cost

       D. Gradient

Ans : Convolution

Q.15 Name the component of a Neural Network where the true value of the input is not observed.

       A. Hidden Layer

       B. Output Layer

       C. Gradient Descent

       D. Activation Function

Ans : Activation Function

Q.16 What is the best Neural Network Model for Temporal Data?

       A. Temporal Neural Networks

       B. Multi Layer Perceptrons

       C. Convolution Neural Networks

       D. Recurrent Neural Network

Ans : Recurrent Neural Network

Q.17 The measure of Difference between two probability distributions is know as ___________.

       A. KL Divergence

       B. Probability Difference

       C. Cost

       D. Error

Ans : KL Divergence

Q.18 All the Visible Layers in a Restricted Boltzmann Machine are connected to each other.

       A. True

       B. False

Ans : True

Q.19 In a Neural Network, all the edges and nodes have the same Weight and Bias values.

       A. True

       B. False

Ans : True

Q.20 What is the difference between the actual output and generated output known as?

       A. Output Difference

       B. Output Modulus

       C. Accuracy

       D. Cost

Ans : Output Difference

Q.21 Process of improving the accuracy of a Neural Network is called __________.

       A. Cross Validation

       B. Random Walk

       C. Forward Propagation

       D. Training

Ans : Training

Q.22 Autoencoders are trained using _________.

       A. They do not require Training

       B. Back Propagation

       C. Feed Forward

       D. Reconstruction

Ans : Back Propagation

Q.23 Data Collected from Survey results is an example of __________.

       A. Structured Data

       B. Unstructured Data

       C. Information

       D. Data

Ans : Structured Data

Q.24 Support Vector Machines, Naive Bayes and Logistic Regression are used for solving ___________ problems.

       A. Classification

       B. Clustering

       C. Time Series

       D. Regression

Ans : Clustering

Q.25 Recurrent Network can input Sequence of Data Points and Produce a Sequence of Output.

       A. True

       B. False

Ans : True

Q.26 Gradient at a given layer is the product of all gradients at the previous layers.

       A. True

       B. False

Ans : True

Q.27 What is the method to overcome the Decay of Information through time in RNN known as?

       A. Gradient Descent

       B. Back Propagation

       C. Activation

       D. Gating

Ans : Back Propagation

Q.28 Why is the Pooling Layer used in a Convolution Neural Network?

       A. Image Sensing

       B. Object Recognition

       C. Dimension Reduction

       D. They are of no use in CNN

Ans : Dimension Reduction

Q.29 Prediction Accuracy of a Neural Network depends on _______________ and ______________.

       A. Activation and Threshold

       B. Linear and Logistic Function

       C. Weight and Bias

       D. Input and Output

Ans : Input and Output

Q.30 ____________ works best for Image Data.

       A. Convolution Networks

       B. Single Layer Perceptrons

       C. Random Forest

       D. AutoEncoders

Ans : AutoEncoders

Q.31 ____________ models are best suited for Recursive Data.

       A. Multi Layer Perceptrons

       B. Convolutional Neural Networks

       C. Recursive Neural Networks

       D. Recursive Neural Tensor Nets

Ans : Recursive Neural Networks

Leave a Comment