{"description":"Evaluate your understanding of the gradient descent algorithm, its purpose, how it works, and how it optimizes model parameters.","questions":[{"answer":"To minimize the loss function by adjusting model parameters","number":1,"options":["To visualize the dataset","To increase the number of features","To minimize the loss function by adjusting model parameters","To normalize data"],"question":"What is the primary goal of the gradient descent algorithm?"},{"answer":"The error or loss in model predictions","number":2,"options":["The learning rate","The number of iterations","The error or loss in model predictions","The input features"],"question":"In gradient descent, imagine you're blindfolded on a mountain trying to reach the bottom by feeling the slope under your feet and taking steps downhill. In this analogy, what does the valley represent?"},{"answer":"The step size of each iteration","number":3,"options":["The total number of features","The direction of the gradient","The step size of each iteration","The final loss value"],"question":"What does the learning rate (alpha) control in gradient descent?"},{"answer":"It may skip over the minimum and fail to converge","number":4,"options":["The algorithm may converge faster","It may skip over the minimum and fail to converge","The model becomes more complex","The gradient becomes zero"],"question":"What happens if the learning rate is too large?"},{"answer":"Using the average of squared differences between predictions and actual values","number":5,"options":["Using cross-entropy","Using the average of absolute errors","Using the average of squared differences between predictions and actual values","Using maximum likelihood estimation"],"question":"How is the loss computed in gradient descent for linear regression?"},{"answer":"The rate of change of loss with respect to model parameters","number":6,"options":["The rate of change of loss with respect to input features","The rate of change of input data","The rate of change of loss with respect to model parameters","Noise in the data"],"question":"In gradient descent for linear regression, L represents the loss (or error), and a and b are the parameters of the model (slope and intercept). What do the partial derivatives \u2202L/\u2202a and \u2202L/\u2202b represent?"},{"answer":"They decrease as the model improves","number":7,"options":["They remain constant","They increase","They decrease as the model improves","They oscillate randomly"],"question":"After several iterations, what happens to the values of \u2202L/\u2202a and \u2202L/\u2202b?"},{"answer":"The loss decreases and eventually levels off","number":8,"options":["The loss remains the same","The loss increases","The loss decreases and eventually levels off","The loss becomes undefined"],"question":"What is the impact of gradient descent on the loss over multiple epochs?"},{"answer":"How well the model predictions align with actual data","number":9,"options":["The change in gradients","The learning rate size","How well the model predictions align with actual data","The number of features in the dataset"],"question":"What does a scatter plot with a fitted regression line help visualize?"},{"answer":"The model has likely found an optimal set of parameters","number":10,"options":["The model parameters are overfitting","The training data is too simple","The model has likely found an optimal set of parameters","The loss function needs to be changed"],"question":"What does a final low value of the loss indicate in gradient descent?"}],"title":"Gradient Descent"}