Unit 6: Artificial neural networks - Practice Quiz

BTY587 — Data Analysis And Simulations 60 Questions
0 Correct 0 Wrong 60 Left
0/60

1 What does the abbreviation ANN stand for?

introduction of deep learning algorithms Easy
A. Analog Network Node
B. Artificial Neural Network
C. Automated Numeric Node
D. Advanced Neural Notation

2 The basic computational unit of an artificial neural network is called a:

introduction of deep learning algorithms Easy
A. Cluster
B. Compiler
C. Neuron (node)
D. Register

3 What is the main purpose of an activation function in a neural network?

introduction of deep learning algorithms Easy
A. To label the dataset
B. To store data permanently
C. To reduce the file size
D. To introduce non-linearity

4 Which term describes a neural network with many hidden layers?

introduction of deep learning algorithms Easy
A. Deep neural network
B. Shallow node network
C. Single-layer network
D. Flat learning network

5 In an ANN, what are the numerical values that are adjusted during training called?

introduction of deep learning algorithms Easy
A. Labels
B. Pixels
C. Weights
D. Tokens

6 Which algorithm is commonly used to train neural networks by updating weights based on error?

introduction of deep learning algorithms Easy
A. Bubble sort
B. Binary search
C. Backpropagation
D. Hashing

7 The function that measures the difference between predicted and actual output is called the:

introduction of deep learning algorithms Easy
A. Search function
B. Loss function
C. Sort function
D. Input function

8 Which optimization technique minimizes the loss function by moving toward its steepest descent?

introduction of deep learning algorithms Easy
A. Linear regression
B. Gradient descent
C. Data cleaning
D. Feature scaling

9 Which type of neural network is most commonly used for image recognition tasks?

types of ANN Easy
A. Hopfield Network
B. Recurrent Neural Network (RNN)
C. Radial Basis Network
D. Convolutional Neural Network (CNN)

10 Which type of neural network is best suited for sequential data such as text or time series?

types of ANN Easy
A. Convolutional Neural Network (CNN)
B. Feedforward Network
C. Recurrent Neural Network (RNN)
D. Autoencoder

11 In a feedforward neural network, information flows in which direction?

types of ANN Easy
A. Only backward, output to input
B. Only forward, input to output
C. Randomly between layers
D. In circular loops

12 Which network type is designed to learn a compressed representation of data for tasks like dimensionality reduction?

types of ANN Easy
A. CNN
B. GAN
C. Perceptron
D. Autoencoder

13 What does GAN stand for in deep learning?

types of ANN Easy
A. General Analog Network
B. Grouped Activation Node
C. Generative Adversarial Network
D. Graph Aligned Network

14 The simplest type of artificial neural network, consisting of a single neuron, is called a:

types of ANN Easy
A. Transformer
B. Boltzmann machine
C. Autoencoder
D. Perceptron

15 Which specialized RNN variant is designed to handle long-term dependencies in sequences?

types of ANN Easy
A. LSTM (Long Short-Term Memory)
B. Perceptron
C. CNN
D. Autoencoder

16 Deep learning is commonly used in healthcare to analyze medical images such as:

case studies for the application of deep learning in biology and health care research Easy
A. Text messages
B. Bank statements
C. Weather reports
D. X-rays and MRI scans

17 In biology, deep learning has been famously applied to predict the 3D structure of:

case studies for the application of deep learning in biology and health care research Easy
A. Proteins
B. Bridges
C. Volcanoes
D. Planets

18 Which healthcare task uses deep learning to identify cancerous cells in tissue slides?

case studies for the application of deep learning in biology and health care research Easy
A. Appointment scheduling
B. Inventory tracking
C. Digital pathology image analysis
D. Payroll processing

19 In genomics, deep learning is used to analyze which type of data?

case studies for the application of deep learning in biology and health care research Easy
A. Stock prices
B. Traffic signals
C. Satellite orbits
D. DNA and gene sequences

20 Which of the following is a healthcare application of deep learning?

case studies for the application of deep learning in biology and health care research Easy
A. Playing background music
B. Predicting disease from patient records
C. Cooking recipes automatically
D. Painting house walls

21 A researcher is designing a network to classify pixel images of handwritten digits while preserving spatial relationships between pixels. Which type of ANN is most appropriate?

types of ANN Medium
A. Recurrent Neural Network (RNN)
B. Single-layer Perceptron
C. Hopfield Network
D. Convolutional Neural Network (CNN)

22 For modeling a time-series of gene expression measured over sequential days, where the value at each step depends on previous steps, which architecture is best suited?

types of ANN Medium
A. Radial Basis Function Network
B. Convolutional Neural Network (CNN)
C. Recurrent Neural Network (RNN)
D. Feedforward Neural Network

23 What is the key limitation of a standard RNN that motivated the development of LSTM networks?

types of ANN Medium
A. Inability to process any sequential data
B. The vanishing gradient problem over long sequences
C. Lack of a nonlinear activation function
D. Excessive memory requirements for small inputs

24 An autoencoder is primarily used for which of the following tasks?

types of ANN Medium
A. Unsupervised learning of compressed data representations
B. Supervised classification of labeled images
C. Sorting data into predefined clusters using labels
D. Reinforcement learning of game strategies

25 In a Generative Adversarial Network (GAN), what are the two competing components?

types of ANN Medium
A. A teacher and a student
B. An encoder and a decoder
C. A forward pass and a backward pass
D. A generator and a discriminator

26 Which statement best distinguishes a feedforward neural network from a recurrent neural network?

types of ANN Medium
A. Both have identical topologies and differ only in activation functions
B. A feedforward network can only handle images
C. A feedforward network has no cycles, while an RNN contains feedback loops that feed outputs back as inputs across time steps to model sequential dependencies
D. A feedforward network uses cycles while an RNN does not

27 Which type of ANN is specifically designed for graph-structured data such as molecular structures or protein interaction networks?

types of ANN Medium
A. Convolutional Neural Network (CNN)
B. Graph Neural Network (GNN)
C. Long Short-Term Memory (LSTM)
D. Multilayer Perceptron (MLP)

28 Why is a nonlinear activation function such as ReLU essential in a deep neural network?

introduction of deep learning algorithms Medium
A. It removes the need for a loss function
B. It guarantees the network will never overfit
C. It reduces the number of parameters needed
D. Without nonlinearity, stacked layers collapse into a single linear transformation

29 The ReLU activation function is defined as which of the following?

introduction of deep learning algorithms Medium
A.
B.
C.
D.

30 During backpropagation, what is being computed and used to update the network weights?

introduction of deep learning algorithms Medium
A. The number of neurons in each hidden layer
B. The raw input values fed forward through the network
C. The random initial weight values
D. The gradient of the loss function with respect to each weight

31 A model achieves 99% accuracy on training data but only 70% on test data. What is this a clear sign of?

introduction of deep learning algorithms Medium
A. Underfitting
B. Vanishing gradients
C. Overfitting
D. A learning rate that is too small

32 Which technique randomly deactivates a fraction of neurons during training to reduce overfitting?

introduction of deep learning algorithms Medium
A. Data augmentation
B. Batch normalization
C. Dropout
D. Gradient clipping

33 In gradient descent, if the learning rate is set too high, what is the most likely consequence?

introduction of deep learning algorithms Medium
A. The gradients become exactly zero
B. The model always converges faster with no downside
C. The network automatically adds more layers
D. The loss may diverge or oscillate instead of converging

34 Which loss function is most appropriate for a multi-class classification problem with a softmax output layer?

introduction of deep learning algorithms Medium
A. Mean squared error
B. Mean absolute error
C. Categorical cross-entropy
D. Hinge loss for regression

35 What is the primary purpose of the softmax function in the output layer of a classification network?

introduction of deep learning algorithms Medium
A. It normalizes the input images before training
B. It reduces the number of output neurons to one
C. It computes the gradient of the loss function directly
D. It converts raw scores into a probability distribution over classes that sums to 1

36 DeepMind's AlphaFold made a major breakthrough in biology by accurately predicting which of the following?

case studies for the application of deep learning in biology and health care research Medium
A. The exact age of a patient from a blood sample
B. The stock prices of pharmaceutical companies
C. The complete DNA sequence of the human genome
D. The 3D structure of proteins from their amino acid sequences

37 In a case study applying CNNs to diagnose diabetic retinopathy, what type of input data is used?

case studies for the application of deep learning in biology and health care research Medium
A. Retinal fundus images
B. Patient audio recordings
C. Genomic sequence text
D. Tabular billing records

38 A hospital deploys a deep learning model to detect tumors in medical scans. Why is high recall (sensitivity) often prioritized over high precision in this context?

case studies for the application of deep learning in biology and health care research Medium
A. Recall requires less computation than precision
B. Missing a true tumor (false negative) can be far more harmful to the patient than a false alarm
C. High recall always guarantees the highest overall accuracy
D. Precision is impossible to compute for medical images

39 When applying deep learning to medical imaging with limited labeled data, which strategy leverages a model pre-trained on a large general image dataset?

case studies for the application of deep learning in biology and health care research Medium
A. Increasing the learning rate
B. Random weight initialization from scratch
C. Removing all hidden layers
D. Transfer learning

40 In genomics, RNNs and CNNs are applied to DNA sequences primarily to accomplish which task?

case studies for the application of deep learning in biology and health care research Medium
A. Identify functional motifs and predict regulatory elements from nucleotide sequences
B. Physically synthesize new DNA molecules in the network
C. Store patient records in a distributed database
D. Replace the need for laboratory sequencing entirely

41 In a Recurrent Neural Network (RNN) trained with Backpropagation Through Time (BPTT), the gradient of the loss with respect to early time steps involves a product of Jacobians. If the largest eigenvalue of the recurrent weight matrix's Jacobian is consistently , what phenomenon dominates?

types of ANN Hard
A. Perfect gradient flow across all time steps
B. Vanishing gradients that hinder learning of long-term dependencies
C. Saturation of the output softmax layer only
D. Exploding gradients that destabilize weight updates

42 A Convolutional Neural Network uses a kernel with stride 1 and no padding on a input. What is the spatial dimension of the resulting feature map?

types of ANN Hard
A.
B.
C.
D.

43 Which architectural property most fundamentally distinguishes an autoencoder used for anomaly detection from a standard feedforward classifier?

types of ANN Hard
A. It applies recurrent connections to model temporal sequences
B. It uses labeled anomaly classes to minimize cross-entropy loss
C. It is trained to reconstruct its input via a bottleneck, using reconstruction error as the signal
D. It relies on max-pooling to downsample spatial features

44 In a Generative Adversarial Network (GAN), the generator and discriminator play a minimax game with value function . At the theoretical global optimum, what value does output for real and generated samples?

types of ANN Hard
A. oscillates between and indefinitely
B. everywhere, since the generator wins
C. everywhere, as the two distributions become indistinguishable
D. for real and for generated samples

45 Why do Long Short-Term Memory (LSTM) networks mitigate the vanishing gradient problem more effectively than vanilla RNNs?

types of ANN Hard
A. They replace all sigmoid activations with ReLU to prevent saturation
B. The cell state provides a nearly linear path allowing gradients to flow largely unattenuated
C. They eliminate recurrent connections entirely, using only feedforward layers
D. They increase the learning rate adaptively at every time step

46 A Radial Basis Function (RBF) network differs from a multilayer perceptron primarily in how its hidden units respond. Which statement is correct?

types of ANN Hard
A. RBF units use global sigmoidal activations identical to MLP hidden units
B. RBF units compute weighted sums followed by softmax normalization
C. RBF units produce localized responses based on distance to a center, giving local approximation
D. RBF networks require at least three hidden layers to function

47 During backpropagation, the ReLU activation has a derivative that is for . What practical failure mode can this cause?

introduction of deep learning algorithms Hard
A. Exploding gradients due to unbounded positive activations
B. Vanishing gradients across all layers regardless of input sign
C. The 'dying ReLU' problem, where neurons output zero and stop updating permanently
D. Guaranteed convergence to the global minimum

48 Adam optimizer combines momentum and RMSProp. Given first moment and second moment , why is bias correction applied in early iterations?

introduction of deep learning algorithms Hard
A. The moment estimates are initialized at zero and are biased toward zero early in training
B. It forces the second moment to equal the first moment
C. It prevents the learning rate from ever decreasing
D. It converts the gradients into probabilities for stability

49 Batch Normalization normalizes layer inputs using mini-batch statistics, then applies learnable parameters and . Why are and necessary?

introduction of deep learning algorithms Hard
A. They fix the mean and variance to and permanently
B. They replace the need for any activation function
C. They compute the gradient of the loss directly
D. They let the network recover representational capacity, including undoing normalization if optimal

50 Dropout with rate is applied during training. At inference time, why must activations be scaled (or weights adjusted)?

introduction of deep learning algorithms Hard
A. To match the expected activation magnitude, since all units are active at test time
B. To zero out half the neurons during inference as well
C. To increase the variance of predictions at test time
D. To convert the network into a recurrent architecture

51 Consider training with cross-entropy loss and a softmax output. Why is the combination of softmax + cross-entropy numerically and computationally preferred over softmax followed by a separate MSE loss for classification?

introduction of deep learning algorithms Hard
A. Its gradient simplifies to , giving stable, well-scaled updates
B. It always produces sparser weight matrices
C. It guarantees the loss surface is globally convex
D. It removes the need for any activation in hidden layers

52 Xavier (Glorot) initialization sets weight variance to . What problem is this designed to prevent?

introduction of deep learning algorithms Hard
A. The optimizer becoming stuck at a saddle point
B. Signal and gradient variance shrinking or growing exponentially across layers
C. Overfitting caused by too many training epochs
D. Class imbalance in the training dataset

53 In transfer learning, a pretrained CNN is adapted to a small medical imaging dataset. Which strategy best reduces overfitting while leveraging learned features?

introduction of deep learning algorithms Hard
A. Freeze the classifier head and train only the input layer
B. Randomly reinitialize all layers and train from scratch
C. Remove all convolutional layers and use only dense layers
D. Freeze early convolutional layers and fine-tune only the later layers and classifier head

54 AlphaFold predicts protein 3D structure from amino acid sequences. Which deep learning innovation was most central to its breakthrough accuracy in CASP14?

case studies for the application of deep learning in biology and health care research Hard
A. A single fully connected network mapping sequence directly to coordinates
B. Unsupervised clustering of protein images without sequence data
C. Attention-based modeling of residue-residue relationships and evolutionary (MSA) information
D. Recurrent networks processing one amino acid at a time without pairwise features

55 A CNN trained to detect diabetic retinopathy achieves 99% accuracy on its training hospital's data but performs poorly at a new clinic. What is the most likely cause?

case studies for the application of deep learning in biology and health care research Hard
A. Domain shift, where imaging equipment and population differences alter the data distribution
B. Cross-entropy loss is inappropriate for image classification
C. The learning rate was set too low during training
D. The model has too few parameters to memorize training data

56 In genomics, a CNN scans DNA sequences to predict transcription factor binding sites. Why are convolutional filters particularly well suited to this task?

case studies for the application of deep learning in biology and health care research Hard
A. They act like position weight matrices detecting motifs regardless of location in the sequence
B. They require the motif to appear at a fixed position to be detected
C. They inherently normalize GC content across the genome
D. They model long-range temporal dependencies better than any other layer

57 When deploying a deep learning model for cancer diagnosis, why is model calibration (not just accuracy) critical?

case studies for the application of deep learning in biology and health care research Hard
A. Predicted probabilities must reflect true likelihoods to support clinical risk decisions
B. Calibration increases the raw accuracy above 100%
C. Calibration replaces the need for a validation dataset
D. Calibration converts the classifier into a regression model

58 A recurrent model (LSTM) is applied to electronic health record (EHR) time series to predict sepsis onset. What key advantage does the LSTM provide over a static feedforward classifier here?

case studies for the application of deep learning in biology and health care research Hard
A. It guarantees interpretability of every prediction to clinicians
B. It captures temporal patterns and irregular dependencies across sequential clinical measurements
C. It eliminates class imbalance between septic and non-septic patients
D. It removes the need to handle missing values in the data

59 In drug discovery, Graph Neural Networks (GNNs) represent molecules as graphs. Why is this representation superior to a fixed-length fingerprint vector for property prediction?

case studies for the application of deep learning in biology and health care research Hard
A. It requires no training data because chemistry is deterministic
B. It ignores molecular topology to focus only on atom counts
C. It converts every molecule into a single scalar automatically
D. Message passing over atoms and bonds learns task-specific structural features directly

60 A U-Net architecture is widely used for biomedical image segmentation. What structural feature most directly enables its precise localization of boundaries?

case studies for the application of deep learning in biology and health care research Hard
A. Purely recurrent connections between encoder and decoder
B. Absence of any downsampling to preserve full resolution throughout
C. Skip connections that fuse high-resolution encoder features with decoder upsampling
D. A single global average pooling layer at the output