Objective 3.2 Optimize the Algorithm (e.g., Structure, Run Time, Tuning Hyperparameters) AIP-210 Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   Why is cross-validation used during hyperparameter tuning?

   • ATo speed up the training process
   • BTo eliminate the need for a test set
   • CTo increase the size of the training dataset
   • DTo evaluate hyperparameter performance on multiple data splits and reduce overfitting
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A team is tuning a model where each training run is expensive, taking several hours. They need to minimize the number of training runs while still finding a good hyperparameter configuration. Which hyperparameter optimization approach is most suitable?

   • AManual tuning, because it allows the team to use their intuition and experience
   • BRandom search, because it is simple and does not require any prior knowledge
   • CBayesian optimization, because it builds a probabilistic model of the objective and selects promising configurations to evaluate
   • DGrid search, because it exhaustively evaluates all combinations and is guaranteed to find the best configuration
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Which structural change could reduce the memory footprint of a machine learning algorithm?

   • AIncreasing the learning rate to speed up convergence
   • BUsing sparse matrix representations instead of dense ones for high-dimensional data
   • CAdding more hidden layers to the neural network
   • DUsing a larger validation set for tuning
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A team is using Optuna to tune hyperparameters for a machine learning model. They want to use an algorithm that intelligently explores the hyperparameter space and reduces the number of trials. Which sampler should they configure in Optuna?

   • ATPESampler, which uses Tree-structured Parzen Estimator to model the objective and suggest promising configurations
   • BCmaEsSampler, which uses a covariance matrix adaptation evolution strategy
   • CRandomSampler, which randomly samples hyperparameters from the search space
   • DGridSampler, which exhaustively evaluates all combinations in the search space
   Reveal Solution Discussion Select an answer first
6. 5application · easy

   A team is tuning a logistic regression model. They are deciding which hyperparameters to tune. Which of the following is a hyperparameter for logistic regression?

   • AThe learned coefficients
   • BThe predicted probabilities
   • CRegularization strength (C)
   • DThe intercept
   Reveal Solution Discussion Select an answer first