Recap Machine Learning

Which of the following statements is true for supervised learning?

• The algorithm learns from input-output pairs to predict outputs for unseen inputs.

• Supervised learning does not require any labeled data.

• Supervised learning can only be used for classification tasks.

• In supervised learning, the output space must be continuous.

Which of the following is a key characteristic of a probabilistic classifier?

• It predicts the class label based solely on distance metrics.

• It provides a probability distribution over class labels for each prediction.

• It uses a decision boundary to separate classes in the feature space.

• It works by minimizing the error between predicted and true values.

Which of the following techniques is commonly used for hyperparameter tuning in machine learning models to improve performance?

• Grid Search

• Cross-validation

• Data augmentation

• Regularization

Which of the following metrics would be most appropriate for evaluating the performance of a regression model?

• Accuracy

• Precision

• Mean Absolute Error (MAE)

• F1-score

Which of the following deep learning architectures is primarily used for image classification tasks due to its ability to capture spatial hierarchies in images?

• Recurrent Neural Networks (RNN)

• Fully Connected Neural Networks (FCNN)

• Support Vector Machines (SVM)

• Convolutional Neural Networks (CNN)

What does generalization refer to in machine learning?

• The model's ability to memorize training data

• The model's ability to perform well on new, unseen data

• The process of increasing training accuracy

• Using more features to improve performance

In machine learning, overfitting means:

• The model works well on every data point

• The model is too simple to learn patterns

• The model learns training data too exactly

• The model always gives the right results

Which of the following is true about a neural network?

• It follows only fixed logic rules to decide

• It is modeled after the human brain system

• It does not require any input data to trainIt works without any training data needed

• It is useful only for working with pictures