Introduction to Machine Learning: Linear Models

What is a key characteristic of Stochastic Gradient Descent (SGD)?

Which of the following describes the advantage of using SGD over traditional Gradient Descent methods?

What is the first step in the Stochastic Gradient Descent algorithm?

What does 'stochastic' refer to in Stochastic Gradient Descent?

In the context of SGD, what is meant by 'mini-batch'?

What happens in the Stochastic Gradient Descent loop when a model converges?

Which of the following statements is true regarding Batch Gradient Descent compared to SGD?

Why is it important to shuffle the training dataset before each iteration in SGD?

What does the loss function quantify in a machine learning model?

Which optimization technique is most commonly used to minimize the loss function?

Which loss function is especially sensitive to outliers in the dataset?

What advantage does Mean Absolute Error (MAE) Loss have over Mean Squared Error (MSE) Loss?

What is the primary characteristic of loss functions in regression tasks?

Which characteristic makes the Mean Squared Error (MSE) Loss suitable for gradient-based optimization?

What does the term 'Huber Loss' refer to in the context of loss functions?

Which loss function calculates the average of the squared differences?

What is a major consequence of using a high learning rate in SGD?

Which theorem states that a neural network with a single hidden layer can approximate any continuous function?

What role does the hidden layer play in a neural network?

Which method can help mitigate the issues of noisy updates in SGD?

What does the output of a neural network's single hidden layer depend on, mathematically?

In the context of the Universal Approximation Theorem, what is required for a neural network to approximate a continuous function?

What can occur if SGD converges too slowly due to a low learning rate?

Which of the following accurately describes the composition of the neural network function?

What is the main criterion for selecting the best hyperplane in a Support Vector Machine?

What happens when a data point lies on the boundary of the separating classes in SVM?

What is a characteristic of SVM in relation to outliers?

What is meant by the term 'soft margin' in SVM?

What is the formula to minimize when a soft margin is applied in SVM?

When data is not linearly separable, what does SVM do?

What does hinge loss represent in the context of SVM?

What is the result of a maximum-margin hyperplane in SVM?

What is a primary function of the activation function in a Perceptron?

What information does the weight of an input provide in a Perceptron?

Which of the following mathematical forms represents the calculation of the weighted sum in a Perceptron?

What is the purpose of the bias in the Perceptron model?

In which scenario would a single-layer Perceptron be used effectively?

What does the output of a Perceptron model indicate when the summed input exceeds a threshold?

Which type of Perceptron model consists of only one layer?

What is added to the weighted sum in a Perceptron to improve its performance?

What is the primary difference between a single-layer perceptron and a multi-layer perceptron?

Which of the following is NOT an advantage of a multi-layer perceptron model?

Which of the following accurately describes the "backward stage" of the multi-layer perceptron training process?

In which of the following scenarios would a multi-layer perceptron model be a suitable choice?

What is a potential drawback of using a multi-layer perceptron model?

Which of the following is NOT a common type of activation function used in a multi-layer perceptron?

What is a common method for evaluating the performance of a multi-layer perceptron model?

What is the significance of the "hidden layers" in a multi-layer perceptron?