Coele1

1. What is the defining characteristic of the training data used in supervised learning?

A) The data is generated randomly by the algorithm.
B) The data is unlabeled, and the model must find patterns on its own
C) The data is always image-based.
D) The data is labeled, meaning each example is paired with a target output.

2. The primary goal of a supervised learning model is to:

A) Discover hidden patterns without any guidance
B) Reduce the dimensionality of the input data for visualization.
C) Generalize from the training data to make accurate predictions on new, unseen data.
D) Memorize the entire training dataset perfectly.

3. In the analogy of a child learning from flashcards, the animal's name on the card represents what component of supervised learning?

P

A) The loss function
B) The label or target output.
C) The model's parameters.
D) The input features.

4. Which of the following tasks is a classic example of a classification problem?

A) Estimating the annual revenue of a company.
B) Diagnosing a tumor as malignant or benign based on medical images.
C) Predicting the selling price of a house based on its features.
D) Forecasting the temperature for tomorrow.

5. A model that predicts the continuous value of a stock price for the next day is solving a:

A) Classification problem
B) Dimensionality reduction problem
C) Clustering problem.
D) Regression problem.

6. What is the core objective of unsupervised learning?

A) To predict a target variable based on labeled examples
B) To achieve perfect accuracy on a held-out test set.
C) To classify emails into spam and non-spam folders
D) To discover the inherent structure, patterns, or relationships within unlabeled data.

7. In the analogy of a child grouping toys without instructions, the act of putting all the cars together is most similar to which unsupervised learning technique?

A) Reinforcement Learning.
B) Regression
C) Clustering
D) Classification

8. Grouping customers based solely on their purchasing behavior, without pre-defined categories, is an application of:

A) Linear Regression, a type of supervised learning.
B) A support vector machine for classification.
C) Clustering, a type of unsupervised learning.
D) Logistic Regression, a type of supervised learning.

9. The main goal of dimensionality reduction techniques like PCA is to:

A) Predict a continuous output variable.
B) Assign categorical labels to each data point.
C) Increase the number of features to improve model accuracy.
D) Reduce the number of features while preserving the most important information in the data.

10. Market basket analysis, which finds rules like "if chips then soda," is a classic example of:
Association rule learning in unsupervised learning.
Classification in supervised learning.
Regression in supervised learning.
Deep learning with neural networks.

A) Classification in supervised learning.
B) Regression in supervised learning.
C) Association rule learning in unsupervised learning.
D) Deep learning with neural networks.

11. Semi-supervised learning is particularly useful in real-world scenarios because:

A) Labeling data is often expensive and time-consuming, so it leverages a small labeled set with a large unlabeled set.
B) It is always more accurate than fully supervised learning.
C) It requires no labeled data at all.
D) It is simpler to implement than unsupervised learning.

12. The fundamental question that a regression model aims to answer is:

A) "Is this pattern anomalous?"
B) "How much?" or "How many?"
C) "What is the underlying group?"
D) "Which category?"

13. The fundamental question that a classification model aims to answer is:

A) "How much?" or "How many?"
B) "How can I reduce the number of features?"
C) "Which category?" or "What class?"
D) "What is the correlation between these variables?"

14. Which algorithm is most directly designed for predicting a continuous target variable?

A) Decision Tree for classification.
B) k-Nearest Neighbors for classification.
C) Linear Regression.
D) Logistic Regression.

15. A model that uses patient data to assign a label of "High," "Medium," or "Low" risk for a disease is performing:

A) Dimensionality reduction.
B) Clustering.
C) Multi-class classification.
D) Regression.

16. In a Decision Tree used for classification, what do the leaf nodes represent?

A) The input features for a new data point.
B) The average value of a continuous target.
C) The final class labels or decisions.
D) The probability of moving to the next node.

17. In a Regression Tree, what is typically represented at the leaf nodes?

A) A continuous value, often the mean of the target values of the training instances that reach the leaf.
B) A random number.
C) The name of the feature used for splitting.
D) A categorical class label.

18. A key strength of Decision Trees is their:

A) Superior performance on all types of data compared to other algorithms.
B) Interpretability; the model's decision-making process is easy to understand and visualize.
C) Guarantee to find the global optimum for any dataset.
D) Immunity to overfitting on noisy datasets.

19. The "kernel trick" used in Support Vector Machines (SVMs) allows them to:

A) Perform linear regression more efficiently.
B) Find a linear separating hyperplane in a high-dimensional feature space, even when the data is not linearly separable in the original space.
C) Grow a tree structure by making sequential decisions.
D) Initialize the weights of a neural network.

20. The "support vectors" in an SVM are the:

A) Data points that are closest to the decision boundary and most critical for defining the optimal hyperplane.
B) The weights of a neural network layer.
C) All data points in the training set.
D) The axes of the original feature space.

21. When comparing Decision Trees and SVMs, a primary advantage of SVMs is:

A) Their inherent resistance to any form of overfitting.
B) Their lower computational cost for very large datasets.
C) Their superior interpretability and simplicity.
D) Their effectiveness in high-dimensional spaces and their ability to model complex, non-linear decision boundaries.

22. The process in supervised learning where a model's parameters are adjusted to minimize the difference between its predictions and the true labels is called:

A) Data preprocessing.
B) Dimensionality reduction.
C) Training or model fitting.
D) Clustering.

23. A key challenge in unsupervised learning is evaluating model performance because:

A) The models are always less accurate than supervised models.
B) There are no ground truth labels to compare the results against.
C) The algorithms are not well-defined.
D) The data is always too small.

24. The task of reducing a 50-dimensional dataset to a 2-dimensional plot for visualization is best accomplished by:

A) An Association rule learning algorithm.
B) A Classification algorithm like Logistic Regression.
C) Dimensionality Reduction techniques like Principal Component Analysis (PCA).
D) A Regression algorithm like Linear Regression.

25. If an e-commerce company wants to automatically group its products into categories without any pre-existing labels, it should use:

A) Clustering, an unsupervised learning method.
B) Regression, a supervised learning method.
C) Classification, a supervised learning method.
D) A neural network for image recognition.

26. The core building block of a neural network is a(n):

A) Decision node in a tree.
B) Principal component.
C) Support vector.
D) Artificial neuron or perceptron, which receives inputs, applies a transformation, and produces an output.

27. In a neural network, the function inside a neuron that determines its output based on the weighted sum of its inputs is called the:

A) Kernel function.
B) Optimization algorithm.
C) Activation function.
D) Loss function.

28. Which of the following is a non-linear activation function crucial for allowing neural networks to learn complex patterns?

A) Rectified Linear Unit (ReLU).
B) The mean squared error function.
C) The identity function (f(x) = x).
D) A constant function.

29. The process of "training" a neural network involves:

A) Manually setting the weights based on expert knowledge.
B) Iteratively adjusting the weights and biases to minimize a loss function.
C) Randomly assigning weights and never changing them.
D) Clustering the input data.

30. Backpropagation is the algorithm used in neural networks to:

A) Initialize the weights before training.
B) Visualize the network's architecture.
C) Perform clustering on the output layer.
D) Efficiently calculate the gradient of the loss function with respect to all the weights in the network, enabling the use of gradient descent.

31. Deep Learning is a subfield of machine learning that primarily uses:

A) Neural networks with many layers (hence "deep").
B) Simple linear regression models.
C) K-means clustering exclusively.
D) Decision trees with a single split.

32. A key advantage of deep neural networks over shallower models is their ability to:

A) Always train faster and with less data.
B) Automatically learn hierarchical feature representations from data.
C) Be perfectly interpretable, like a decision tree.
D) Operate without any need for data preprocessing.

33. Convolutional Neural Networks (CNNs) are particularly well-suited for tasks involving:

A) Tabular data with many categorical features.
B) Image data, due to their architecture which exploits spatial locality.
C) Text data and natural language processing.
D) Unsupervised clustering of audio signals.

34. The "convolution" operation in a CNN is designed to:

A) Perform the final classification.
B) Flatten the input into a single vector.
C) Initialize the weights of the network.
D) Detect local features (like edges or textures) in the input by applying a set of learnable filters.

35. Recurrent Neural Networks (RNNs) are designed to handle:

A) Only image data.
B) Static, non-temporal data.
C) Sequential data, like time series or text, due to their internal "memory" of previous inputs.
D) Independent and identically distributed (IID) data points.

36. The "vanishing gradient" problem in deep networks refers to:

A) The loss function reaching a perfect value of zero.
B) The model overfitting to the training data.
C) The gradients becoming too large and causing numerical instability.
D) The gradients becoming exceedingly small as they are backpropagated through many layers, which can halt learning in early layers.

37. The "training set" is used to:

A) Fit the model's parameters (e.g., the weights in a neural network).
B) Deploy the model in a production environment.
C) Provide an unbiased evaluation of a final model's performance.
D) Tune the model's hyperparameters.

38. The "validation set" is primarily used for:

A) Data preprocessing and cleaning.
B) The initial training of the model's weights.
C) Tuning hyperparameters and making decisions about the model architecture during development.
D) The final, unbiased assessment of the model's generalization error.

39. The "test set" should be:

A) Ignored in the machine learning pipeline.
B) Used repeatedly to tune the model's hyperparameters.
C) Used as part of the training data to improve accuracy.
D) Used only once, for a final evaluation of the model's performance on unseen data after model development is complete.

40. Overfitting occurs when a model:

A) Is evaluated using the training set instead of a test set.
B) Is too simple to capture the trends in the data.
C) Learns the training data too well, including its noise and outliers, and performs poorly on new, unseen data.
D) Fails to learn the underlying pattern in the training data.

41. A common technique to reduce overfitting in neural networks is:

A) Training for more epochs without any checks.
B) Using a smaller training dataset.
C) Dropout, which randomly ignores a subset of neurons during training.
D) Increasing the model's capacity by adding more layers.

42. The "bias" of a model refers to:

A) The weights connecting the input layer to the hidden layer.
B) The error from erroneous assumptions in the learning algorithm, leading to underfitting.
C) The activation function used in the output layer.
D) The error from sensitivity to small fluctuations in the training set, leading to overfitting.

43. The "variance" of a model refers to:

A) The speed at which the model trains.
B) The error from erroneous assumptions in the learning algorithm, leading to underfitting.
C) The error from sensitivity to small fluctuations in the training set, leading to overfitting.
D) The intercept term in a linear regression model.

44. The "bias-variance tradeoff" implies that:

A) Only bias is important for model performance.
B) Decreasing bias will typically increase variance, and vice versa. The goal is to find a balance.
C) Only variance is important for model performance.
D) Bias and variance can be minimized to zero simultaneously.

45. A learning curve that shows high training accuracy but low validation accuracy is a classic sign of:

A) A well-generalized model.
B) Underfitting.
C) Perfect model performance.
D) Overfitting.

46. In a neural network, the "loss function" (or cost function) measures:

A) The accuracy on the test set.
B) How well the model is performing on the training data; it's the quantity we want to minimize during training.
C) The speed of the backpropagation algorithm.
D) The number of layers in the network.

47. Gradient Descent is an optimization algorithm that:

A) Guarantees finding the global minimum for any loss function.
B) Iteratively adjusts parameters in the direction that reduces the loss function.
C) Is only used for unsupervised learning.
D) Randomly searches the parameter space for a good solution.

48. The "learning rate" in gradient descent controls:

A) The size of the step taken during each parameter update. A rate that is too high can cause divergence, while one that is too low can make training slow.
B) The amount of training data used in each epoch.
C) The number of layers in a neural network.
D) The activation function for the output layer.

49. "Epoch" in neural network training refers to:

A) The final evaluation on the test set.
B) A type of regularization technique.
C) One complete pass of the entire training dataset through the learning algorithm.
D) The processing of a single training example.

50. "Batch Size" in neural network training refers to:

A) The total number of examples in the training set.
B) The number of layers in the network.
C) The number of training examples used in one forward/backward pass before the model's parameters are updated.
D) The number of validation examples.

51. "Stochastic Gradient Descent" (SGD) uses a batch size of:

A) A random number between 1 and 100.
B) 1, meaning the parameters are updated after each individual training example.
C) The entire training set.
D) Exactly 50% of the training set.

Created with That Quiz — a math test site for students of all grade levels.