NEW
Font size
WorksheetsPython Mastery
Total questions: 12
Worksheet time: 7mins
Which of the following best describes deep learning?
A subset of machine learning focused on algorithms with multiple layers
A type of unsupervised learning technique
A traditional approach to programming
A method for analyzing structured data only
What is the primary advantage of convolutional neural networks (CNNs) in image processing?
Ability to handle only one layer of neurons
Capability to automatically learn hierarchical features from raw data
Reliance on handcrafted features for image analysis
Limited applicability to tasks beyond image processing
What role do pooling layers play in convolutional neural networks (CNNs)?
Increasing the size of feature maps
Downsampling the feature maps to reduce dimensionality
Introducing non-linearity to the network
Adding noise to the input data
Which of the following is the widely used library for deep learning framework?
NumPy
Matplotlib
TensorFlow
Scikit-Learn
Which of the following is a popular loss function used in the binary classification task for a deep learning model?
Huber Loss
Cross-Entropy Loss
Mean Squared Error (MSE)
Mean Absolute Error (MAE)
What is the primary purpose of modifying capacity in a dense neural network?
To reduce overfitting
To reduce underfitting
To decrease the model's complexity
To balance between underfitting and overfitting the training data
Which type of layers is not part of the Convolutional Neural Network architecture?
Input layer
Middle layer
Pooling layer
Convolutional layer
There are 2 types of pooling layers which are commonly used i.e.
layers
filter layer and combined layer
average pooling layer and Max pooling layer
max pooling layer and minimum pooling layer
Grayscale and RGB images can be represented by matrix shape of
(n x m x 3), (m x m x 0)
(n x m x 1), (n x m x 3)
(m x m x 1), (m x m x 2)
(m x m x 3), (m x m x 2)
Pre-trained part in the Covnets model refer to
base
body
head
tail
The following are advantages of pre-trained model, EXCEPT
Faster optimization
Higher performance
Require less data for training
Less accurate than models trained from scratch.
Regarding transfer learning.
Allows models to be optimized quickly, especially when starting from scratch
Requires less labeled data for fine-tuning
Focuses on memorizing specific patterns from the training data
Pre-trained models cannot be fine-tuned for specific tasks
