{"id":7,"title":"Инно ИС для подготовки","description":null,"category":"Programming","source_filename":"Инно ИС для подготовки.txt","questions":[{"text":"Which statement correctly describes the difference between classification and regression?","explanation":null,"position":1,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Classification predicts continuous values, while regression predicts categories","is_correct":false,"position":1,"images":[]},{"text":"Classification predicts categories, while regression predicts continuous values","is_correct":true,"position":2,"images":[]},{"text":"Both classification and regression always predict only binary labels","is_correct":false,"position":3,"images":[]},{"text":"Regression is used only for unsupervised learning","is_correct":false,"position":4,"images":[]}]},{"text":"Which of the following is an example of a regression task?","explanation":null,"position":2,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Detecting whether an email is spam or not","is_correct":false,"position":1,"images":[]},{"text":"Classifying a tumor as benign or malignant","is_correct":false,"position":2,"images":[]},{"text":"Predicting the price of a house","is_correct":true,"position":3,"images":[]},{"text":"Grouping customers into clusters","is_correct":false,"position":4,"images":[]}]},{"text":"Which of the following is an example of a classification task?","explanation":null,"position":3,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Predicting tomorrow's temperature","is_correct":false,"position":1,"images":[]},{"text":"Predicting whether a patient has a disease or not","is_correct":true,"position":2,"images":[]},{"text":"Estimating the price of a car","is_correct":false,"position":3,"images":[]},{"text":"Reducing the number of features in a dataset","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main characteristic of supervised machine learning?","explanation":null,"position":4,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The model learns without any target labels","is_correct":false,"position":1,"images":[]},{"text":"The model learns from input data and known target labels","is_correct":true,"position":2,"images":[]},{"text":"The model only visualizes data","is_correct":false,"position":3,"images":[]},{"text":"The model randomly assigns clusters","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main characteristic of unsupervised machine learning?","explanation":null,"position":5,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It requires labeled training examples","is_correct":false,"position":1,"images":[]},{"text":"It always predicts numerical target values","is_correct":false,"position":2,"images":[]},{"text":"It finds patterns in data without target labels","is_correct":true,"position":3,"images":[]},{"text":"It can only be used for classification","is_correct":false,"position":4,"images":[]}]},{"text":"What is the purpose of training data in machine learning?","explanation":null,"position":6,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To teach the model patterns from examples","is_correct":true,"position":1,"images":[]},{"text":"To test the model only after deployment","is_correct":false,"position":2,"images":[]},{"text":"To remove all features from the dataset","is_correct":false,"position":3,"images":[]},{"text":"To manually write all prediction rules","is_correct":false,"position":4,"images":[]}]},{"text":"Why is test data used in machine learning?","explanation":null,"position":7,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To train the model parameters","is_correct":false,"position":1,"images":[]},{"text":"To evaluate how well the model performs on unseen data","is_correct":true,"position":2,"images":[]},{"text":"To increase the number of training examples","is_correct":false,"position":3,"images":[]},{"text":"To replace feature scaling","is_correct":false,"position":4,"images":[]}]},{"text":"What does generalization mean in machine learning?","explanation":null,"position":8,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Memorizing all training examples perfectly","is_correct":false,"position":1,"images":[]},{"text":"Performing well on new, unseen data","is_correct":true,"position":2,"images":[]},{"text":"Using only one feature for prediction","is_correct":false,"position":3,"images":[]},{"text":"Removing the test set from the workflow","is_correct":false,"position":4,"images":[]}]},{"text":"What is overfitting?","explanation":null,"position":9,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A model is too simple to learn patterns","is_correct":false,"position":1,"images":[]},{"text":"A model fits training data too closely and performs worse on new data","is_correct":true,"position":2,"images":[]},{"text":"A model has no parameters","is_correct":false,"position":3,"images":[]},{"text":"A model works only with unlabeled data","is_correct":false,"position":4,"images":[]}]},{"text":"What is underfitting?","explanation":null,"position":10,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A model is too simple and fails to capture important patterns","is_correct":true,"position":1,"images":[]},{"text":"A model learns training data too well","is_correct":false,"position":2,"images":[]},{"text":"A model has too many decision rules","is_correct":false,"position":3,"images":[]},{"text":"A model performs perfectly on training data","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main idea of the K-Nearest Neighbors algorithm?","explanation":null,"position":11,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It builds a sequence of decision trees","is_correct":false,"position":1,"images":[]},{"text":"It predicts using the closest training examples","is_correct":true,"position":2,"images":[]},{"text":"It calculates principal components","is_correct":false,"position":3,"images":[]},{"text":"It creates random clusters without distances","is_correct":false,"position":4,"images":[]}]},{"text":"How does KNN usually make a classification prediction?","explanation":null,"position":12,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"By averaging all features in the dataset","is_correct":false,"position":1,"images":[]},{"text":"By majority vote among the nearest neighbors","is_correct":true,"position":2,"images":[]},{"text":"By selecting the farthest neighbor","is_correct":false,"position":3,"images":[]},{"text":"By calculating a linear equation","is_correct":false,"position":4,"images":[]}]},{"text":"How does KNN usually make a regression prediction?","explanation":null,"position":13,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"By choosing the most frequent class","is_correct":false,"position":1,"images":[]},{"text":"By averaging the target values of nearest neighbors","is_correct":true,"position":2,"images":[]},{"text":"By creating a dendrogram","is_correct":false,"position":3,"images":[]},{"text":"By applying logistic regression","is_correct":false,"position":4,"images":[]}]},{"text":"What can happen when K=1 in KNN?","explanation":null,"position":14,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The model can become sensitive to noise and overfit","is_correct":true,"position":1,"images":[]},{"text":"The model always becomes smoother","is_correct":false,"position":2,"images":[]},{"text":"The model ignores the training data","is_correct":false,"position":3,"images":[]},{"text":"The model becomes a Naive Bayes classifier","is_correct":false,"position":4,"images":[]}]},{"text":"What usually happens when the number of neighbors in KNN increases?","explanation":null,"position":15,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The model becomes more complex and unstable","is_correct":false,"position":1,"images":[]},{"text":"The model becomes smoother and simpler","is_correct":true,"position":2,"images":[]},{"text":"The model stops using distance","is_correct":false,"position":3,"images":[]},{"text":"The model becomes a clustering algorithm","is_correct":false,"position":4,"images":[]}]},{"text":"Why is distance important in KNN?","explanation":null,"position":16,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"KNN does not use distances","is_correct":false,"position":1,"images":[]},{"text":"The nearest examples are selected based on distance","is_correct":true,"position":2,"images":[]},{"text":"Distance is used only in decision trees","is_correct":false,"position":3,"images":[]},{"text":"Distance replaces the target variable","is_correct":false,"position":4,"images":[]}]},{"text":"Why is feature scaling important for KNN?","explanation":null,"position":17,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"KNN works only with unscaled categorical data","is_correct":false,"position":1,"images":[]},{"text":"Features with larger scales can dominate distance calculations","is_correct":true,"position":2,"images":[]},{"text":"Scaling removes the need for training data","is_correct":false,"position":3,"images":[]},{"text":"Scaling converts regression into classification","is_correct":false,"position":4,"images":[]}]},{"text":"Which is an advantage of KNN?","explanation":null,"position":18,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It never requires data preprocessing","is_correct":false,"position":1,"images":[]},{"text":"It is simple and easy to interpret","is_correct":true,"position":2,"images":[]},{"text":"It always works best with thousands of features","is_correct":false,"position":3,"images":[]},{"text":"It does not require a value of K","is_correct":false,"position":4,"images":[]}]},{"text":"Which is a limitation of KNN?","explanation":null,"position":19,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It cannot use numerical data","is_correct":false,"position":1,"images":[]},{"text":"It may work poorly with very high-dimensional data","is_correct":true,"position":2,"images":[]},{"text":"It cannot be used for regression","is_correct":false,"position":3,"images":[]},{"text":"It does not depend on training examples","is_correct":false,"position":4,"images":[]}]},{"text":"What does a linear regression model use to make predictions?","explanation":null,"position":20,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only random cluster labels","is_correct":false,"position":1,"images":[]},{"text":"A weighted sum of features plus an intercept","is_correct":true,"position":2,"images":[]},{"text":"Only the nearest neighbor","is_correct":false,"position":3,"images":[]},{"text":"A dendrogram of classes","is_correct":false,"position":4,"images":[]}]},{"text":"In a linear model, what do weights represent?","explanation":null,"position":21,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The influence of features on the prediction","is_correct":true,"position":1,"images":[]},{"text":"The number of rows in the dataset","is_correct":false,"position":2,"images":[]},{"text":"The number of clusters","is_correct":false,"position":3,"images":[]},{"text":"The size of the test set","is_correct":false,"position":4,"images":[]}]},{"text":"For one input feature, the prediction of a linear regression model is represented as:","explanation":null,"position":22,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A circle","is_correct":false,"position":1,"images":[]},{"text":"A straight line","is_correct":true,"position":2,"images":[]},{"text":"A decision tree","is_correct":false,"position":3,"images":[]},{"text":"A cluster center only","is_correct":false,"position":4,"images":[]}]},{"text":"What does ordinary least squares aim to do in linear regression?","explanation":null,"position":23,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Maximize classification error","is_correct":false,"position":1,"images":[]},{"text":"Find coefficients that minimize prediction error on training data","is_correct":true,"position":2,"images":[]},{"text":"Create clusters of equal size","is_correct":false,"position":3,"images":[]},{"text":"Remove all numerical features","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main purpose of Ridge Regression?","explanation":null,"position":24,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To increase the complexity of a model without control","is_correct":false,"position":1,"images":[]},{"text":"To reduce overfitting by shrinking coefficients","is_correct":true,"position":2,"images":[]},{"text":"To create cluster labels","is_correct":false,"position":3,"images":[]},{"text":"To remove the intercept from the model","is_correct":false,"position":4,"images":[]}]},{"text":"What happens when the alpha parameter in Ridge Regression increases?","explanation":null,"position":25,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Coefficients become completely random","is_correct":false,"position":1,"images":[]},{"text":"Coefficients are pushed closer to zero","is_correct":true,"position":2,"images":[]},{"text":"The number of classes increases","is_correct":false,"position":3,"images":[]},{"text":"The model becomes K-Means","is_correct":false,"position":4,"images":[]}]},{"text":"How can Ridge Regression improve generalization?","explanation":null,"position":26,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"By using no training data","is_correct":false,"position":1,"images":[]},{"text":"By limiting model complexity through regularization","is_correct":true,"position":2,"images":[]},{"text":"By creating more target labels","is_correct":false,"position":3,"images":[]},{"text":"By fitting the test set separately","is_correct":false,"position":4,"images":[]}]},{"text":"What is a key property of Lasso Regression?","explanation":null,"position":27,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It can make some coefficients exactly zero","is_correct":true,"position":1,"images":[]},{"text":"It never performs feature selection","is_correct":false,"position":2,"images":[]},{"text":"It is used only for clustering","is_correct":false,"position":3,"images":[]},{"text":"It always uses all features equally","is_correct":false,"position":4,"images":[]}]},{"text":"Why can Lasso be useful when there are many features?","explanation":null,"position":28,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It increases all coefficients equally","is_correct":false,"position":1,"images":[]},{"text":"It can select only some important features","is_correct":true,"position":2,"images":[]},{"text":"It removes the target variable","is_correct":false,"position":3,"images":[]},{"text":"It converts numerical values into text","is_correct":false,"position":4,"images":[]}]},{"text":"Which statement is correct about Ridge and Lasso?","explanation":null,"position":29,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Both always set all coefficients to zero","is_correct":false,"position":1,"images":[]},{"text":"Ridge shrinks coefficients, while Lasso can set some coefficients to exactly zero","is_correct":true,"position":2,"images":[]},{"text":"Ridge is only for clustering, while Lasso is only for PCA","is_correct":false,"position":3,"images":[]},{"text":"Neither method uses regularization","is_correct":false,"position":4,"images":[]}]},{"text":"Logistic Regression is mainly used for which type of task?","explanation":null,"position":30,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Classification","is_correct":true,"position":1,"images":[]},{"text":"Clustering","is_correct":false,"position":2,"images":[]},{"text":"Dimensionality reduction","is_correct":false,"position":3,"images":[]},{"text":"Image reconstruction only","is_correct":false,"position":4,"images":[]}]},{"text":"In linear classification, how is a class prediction commonly made?","explanation":null,"position":31,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"By calculating only the average of all labels","is_correct":false,"position":1,"images":[]},{"text":"By applying a threshold to a linear function of features","is_correct":true,"position":2,"images":[]},{"text":"By using a dendrogram","is_correct":false,"position":3,"images":[]},{"text":"By randomly choosing the nearest cluster","is_correct":false,"position":4,"images":[]}]},{"text":"LinearSVC is an example of which type of model?","explanation":null,"position":32,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Unsupervised clustering model","is_correct":false,"position":1,"images":[]},{"text":"Linear classification model","is_correct":true,"position":2,"images":[]},{"text":"Dimensionality reduction model","is_correct":false,"position":3,"images":[]},{"text":"Feature scaling model","is_correct":false,"position":4,"images":[]}]},{"text":"In Logistic Regression and LinearSVC, what does parameter C control?","explanation":null,"position":33,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The number of clusters","is_correct":false,"position":1,"images":[]},{"text":"The strength of regularization","is_correct":true,"position":2,"images":[]},{"text":"The number of principal components","is_correct":false,"position":3,"images":[]},{"text":"The size of the dataset","is_correct":false,"position":4,"images":[]}]},{"text":"What does a small value of C usually mean in linear classification models?","explanation":null,"position":34,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Stronger regularization and a simpler model","is_correct":true,"position":1,"images":[]},{"text":"Weaker regularization and a more complex model","is_correct":false,"position":2,"images":[]},{"text":"More clusters","is_correct":false,"position":3,"images":[]},{"text":"No training process","is_correct":false,"position":4,"images":[]}]},{"text":"What does a large value of C usually mean in linear classification models?","explanation":null,"position":35,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Stronger regularization","is_correct":false,"position":1,"images":[]},{"text":"Weaker regularization and a more complex model","is_correct":true,"position":2,"images":[]},{"text":"No use of features","is_correct":false,"position":3,"images":[]},{"text":"Automatic dimensionality reduction","is_correct":false,"position":4,"images":[]}]},{"text":"What is multiclass classification?","explanation":null,"position":36,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Predicting only yes or no labels","is_correct":false,"position":1,"images":[]},{"text":"Predicting one class from more than two possible classes","is_correct":true,"position":2,"images":[]},{"text":"Predicting a continuous number","is_correct":false,"position":3,"images":[]},{"text":"Grouping unlabeled data only","is_correct":false,"position":4,"images":[]}]},{"text":"Which task is an example of binary classification?","explanation":null,"position":37,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Predicting house price","is_correct":false,"position":1,"images":[]},{"text":"Classifying an email as spam or not spam","is_correct":true,"position":2,"images":[]},{"text":"Grouping users into five clusters","is_correct":false,"position":3,"images":[]},{"text":"Reducing 30 features to 2 components","is_correct":false,"position":4,"images":[]}]},{"text":"Which is an advantage of linear models?","explanation":null,"position":38,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They are fast in training and prediction","is_correct":true,"position":1,"images":[]},{"text":"They always outperform all other models","is_correct":false,"position":2,"images":[]},{"text":"They do not need input features","is_correct":false,"position":3,"images":[]},{"text":"They work only for image data","is_correct":false,"position":4,"images":[]}]},{"text":"Why are SGDClassifier and SGDRegressor useful?","explanation":null,"position":39,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They are only for small datasets","is_correct":false,"position":1,"images":[]},{"text":"They are scalable versions of linear models","is_correct":true,"position":2,"images":[]},{"text":"They do not require optimization","is_correct":false,"position":3,"images":[]},{"text":"They replace all clustering algorithms","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main idea of Naive Bayes classifiers?","explanation":null,"position":40,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They consider each feature separately under a simplifying independence assumption","is_correct":true,"position":1,"images":[]},{"text":"They always build many decision trees","is_correct":false,"position":2,"images":[]},{"text":"They require no training data","is_correct":false,"position":3,"images":[]},{"text":"They only perform dimensionality reduction","is_correct":false,"position":4,"images":[]}]},{"text":"GaussianNB is most suitable for which type of data?","explanation":null,"position":41,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only binary data","is_correct":false,"position":1,"images":[]},{"text":"Continuous numerical data","is_correct":true,"position":2,"images":[]},{"text":"Only image pixels with no labels","is_correct":false,"position":3,"images":[]},{"text":"Only cluster labels","is_correct":false,"position":4,"images":[]}]},{"text":"BernoulliNB is commonly suitable for which type of data?","explanation":null,"position":42,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Binary features","is_correct":true,"position":1,"images":[]},{"text":"Continuous regression targets","is_correct":false,"position":2,"images":[]},{"text":"Hierarchical clusters","is_correct":false,"position":3,"images":[]},{"text":"Principal components only","is_correct":false,"position":4,"images":[]}]},{"text":"MultinomialNB is often used for:","explanation":null,"position":43,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Continuous sensor values only","is_correct":false,"position":1,"images":[]},{"text":"Count-based or discrete data such as text word counts","is_correct":true,"position":2,"images":[]},{"text":"Decision tree pruning","is_correct":false,"position":3,"images":[]},{"text":"K-Means initialization","is_correct":false,"position":4,"images":[]}]},{"text":"What is a decision tree?","explanation":null,"position":44,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A model that makes predictions through a sequence of decision rules","is_correct":true,"position":1,"images":[]},{"text":"A model that only scales data","is_correct":false,"position":2,"images":[]},{"text":"A model that finds only principal components","is_correct":false,"position":3,"images":[]},{"text":"A model that cannot be visualized","is_correct":false,"position":4,"images":[]}]},{"text":"DecisionTreeClassifier is used for:","explanation":null,"position":45,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Classification tasks","is_correct":true,"position":1,"images":[]},{"text":"Only clustering tasks","is_correct":false,"position":2,"images":[]},{"text":"Only PCA visualization","is_correct":false,"position":3,"images":[]},{"text":"Only feature scaling","is_correct":false,"position":4,"images":[]}]},{"text":"DecisionTreeRegressor is used for:","explanation":null,"position":46,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Class label prediction only","is_correct":false,"position":1,"images":[]},{"text":"Predicting continuous target values","is_correct":true,"position":2,"images":[]},{"text":"Scaling features to 0 and 1","is_correct":false,"position":3,"images":[]},{"text":"Calculating Naive Bayes probabilities","is_correct":false,"position":4,"images":[]}]},{"text":"Why can decision trees overfit easily?","explanation":null,"position":47,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They are always too simple","is_correct":false,"position":1,"images":[]},{"text":"They can create very detailed rules for training data","is_correct":true,"position":2,"images":[]},{"text":"They cannot split data","is_correct":false,"position":3,"images":[]},{"text":"They do not use target values","is_correct":false,"position":4,"images":[]}]},{"text":"What is pre-pruning in decision trees?","explanation":null,"position":48,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Stopping tree growth early to reduce overfitting","is_correct":true,"position":1,"images":[]},{"text":"Removing the target variable","is_correct":false,"position":2,"images":[]},{"text":"Scaling data before training","is_correct":false,"position":3,"images":[]},{"text":"Increasing the number of trees","is_correct":false,"position":4,"images":[]}]},{"text":"What is post-pruning in decision trees?","explanation":null,"position":49,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Training a tree without data","is_correct":false,"position":1,"images":[]},{"text":"Removing or reducing uninformative nodes after building the tree","is_correct":true,"position":2,"images":[]},{"text":"Converting a tree into PCA","is_correct":false,"position":3,"images":[]},{"text":"Choosing the nearest neighbor","is_correct":false,"position":4,"images":[]}]},{"text":"What does feature importance show in tree-based models?","explanation":null,"position":50,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The number of missing values only","is_correct":false,"position":1,"images":[]},{"text":"How useful each feature is for prediction","is_correct":true,"position":2,"images":[]},{"text":"The number of clusters","is_correct":false,"position":3,"images":[]},{"text":"The number of test samples","is_correct":false,"position":4,"images":[]}]},{"text":"What is an ensemble method?","explanation":null,"position":51,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A method that uses only one simple model","is_correct":false,"position":1,"images":[]},{"text":"A method that combines multiple models to improve performance","is_correct":true,"position":2,"images":[]},{"text":"A method used only for scaling","is_correct":false,"position":3,"images":[]},{"text":"A method that removes training data","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main idea of Random Forest?","explanation":null,"position":52,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To use one linear model","is_correct":false,"position":1,"images":[]},{"text":"To combine many decision trees","is_correct":true,"position":2,"images":[]},{"text":"To reduce data to two dimensions","is_correct":false,"position":3,"images":[]},{"text":"To calculate only Euclidean distance","is_correct":false,"position":4,"images":[]}]},{"text":"Why is Random Forest often more stable than a single decision tree?","explanation":null,"position":53,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It ignores all weak features","is_correct":false,"position":1,"images":[]},{"text":"It averages or combines predictions from many trees","is_correct":true,"position":2,"images":[]},{"text":"It uses no training labels","is_correct":false,"position":3,"images":[]},{"text":"It performs only clustering","is_correct":false,"position":4,"images":[]}]},{"text":"What does the n_estimators parameter usually define in Random Forest?","explanation":null,"position":54,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The number of features","is_correct":false,"position":1,"images":[]},{"text":"The number of trees","is_correct":true,"position":2,"images":[]},{"text":"The number of classes","is_correct":false,"position":3,"images":[]},{"text":"The number of clusters","is_correct":false,"position":4,"images":[]}]},{"text":"What is the purpose of n_jobs=-1 in Random Forest?","explanation":null,"position":55,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To use only one CPU core","is_correct":false,"position":1,"images":[]},{"text":"To use all available CPU cores for parallel computation","is_correct":true,"position":2,"images":[]},{"text":"To remove all trees","is_correct":false,"position":3,"images":[]},{"text":"To disable training","is_correct":false,"position":4,"images":[]}]},{"text":"Which is a limitation of Random Forest?","explanation":null,"position":56,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It cannot handle classification tasks","is_correct":false,"position":1,"images":[]},{"text":"It may work poorly on very high-dimensional sparse data such as text","is_correct":true,"position":2,"images":[]},{"text":"It always requires no memory","is_correct":false,"position":3,"images":[]},{"text":"It cannot estimate feature importance","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main idea of Gradient Boosting?","explanation":null,"position":57,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Build many independent trees without order","is_correct":false,"position":1,"images":[]},{"text":"Build trees sequentially so each new tree corrects previous errors","is_correct":true,"position":2,"images":[]},{"text":"Use only the closest neighbor","is_correct":false,"position":3,"images":[]},{"text":"Scale training and test data separately","is_correct":false,"position":4,"images":[]}]},{"text":"How does Gradient Boosting differ from Random Forest?","explanation":null,"position":58,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Gradient Boosting does not use trees","is_correct":false,"position":1,"images":[]},{"text":"Gradient Boosting builds trees sequentially, while Random Forest builds many trees more independently","is_correct":true,"position":2,"images":[]},{"text":"Random Forest is always unsupervised","is_correct":false,"position":3,"images":[]},{"text":"Gradient Boosting is only used for PCA","is_correct":false,"position":4,"images":[]}]},{"text":"Why does Gradient Boosting require careful parameter tuning?","explanation":null,"position":59,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Because model complexity and overfitting must be controlled","is_correct":true,"position":1,"images":[]},{"text":"Because it has no parameters","is_correct":false,"position":2,"images":[]},{"text":"Because it cannot learn from labels","is_correct":false,"position":3,"images":[]},{"text":"Because it only works with binary features","is_correct":false,"position":4,"images":[]}]},{"text":"What is the purpose of data preprocessing in machine learning?","explanation":null,"position":60,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To prepare data for better model training and evaluation","is_correct":true,"position":1,"images":[]},{"text":"To delete all target labels","is_correct":false,"position":2,"images":[]},{"text":"To replace machine learning algorithms","is_correct":false,"position":3,"images":[]},{"text":"To create random answers","is_correct":false,"position":4,"images":[]}]},{"text":"What does feature scaling do?","explanation":null,"position":61,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Removes the need for features","is_correct":false,"position":1,"images":[]},{"text":"Changes feature values to comparable ranges","is_correct":true,"position":2,"images":[]},{"text":"Creates new class labels","is_correct":false,"position":3,"images":[]},{"text":"Builds a decision tree automatically","is_correct":false,"position":4,"images":[]}]},{"text":"What does MinMaxScaler usually do?","explanation":null,"position":62,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Converts all values into class labels","is_correct":false,"position":1,"images":[]},{"text":"Scales features to a fixed range, often from 0 to 1","is_correct":true,"position":2,"images":[]},{"text":"Removes outliers automatically","is_correct":false,"position":3,"images":[]},{"text":"Creates clusters","is_correct":false,"position":4,"images":[]}]},{"text":"What does StandardScaler usually do?","explanation":null,"position":63,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Scales features using only the test set","is_correct":false,"position":1,"images":[]},{"text":"Transforms features to have zero mean and unit variance","is_correct":true,"position":2,"images":[]},{"text":"Creates decision boundaries","is_correct":false,"position":3,"images":[]},{"text":"Converts regression into classification","is_correct":false,"position":4,"images":[]}]},{"text":"Why should a scaler be fitted only on the training data?","explanation":null,"position":64,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To avoid using information from the test data","is_correct":true,"position":1,"images":[]},{"text":"To make the test data easier to memorize","is_correct":false,"position":2,"images":[]},{"text":"To remove the need for a model","is_correct":false,"position":3,"images":[]},{"text":"To force all test values to be zero","is_correct":false,"position":4,"images":[]}]},{"text":"Why is it wrong to fit a scaler separately on the test data?","explanation":null,"position":65,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It improves evaluation fairness","is_correct":false,"position":1,"images":[]},{"text":"It changes the test data using its own statistics and makes evaluation inconsistent","is_correct":true,"position":2,"images":[]},{"text":"It is required for all algorithms","is_correct":false,"position":3,"images":[]},{"text":"It increases the number of classes","is_correct":false,"position":4,"images":[]}]},{"text":"Why can scaling improve SVM performance?","explanation":null,"position":66,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"SVM ignores feature values","is_correct":false,"position":1,"images":[]},{"text":"Because SVM can be sensitive to feature scales","is_correct":true,"position":2,"images":[]},{"text":"Because scaling creates more labels","is_correct":false,"position":3,"images":[]},{"text":"Because SVM is a clustering algorithm","is_correct":false,"position":4,"images":[]}]},{"text":"What is dimensionality reduction?","explanation":null,"position":67,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Increasing the number of features","is_correct":false,"position":1,"images":[]},{"text":"Reducing the number of features while preserving important information","is_correct":true,"position":2,"images":[]},{"text":"Removing all data points","is_correct":false,"position":3,"images":[]},{"text":"Changing classification into regression","is_correct":false,"position":4,"images":[]}]},{"text":"What is PCA mainly used for?","explanation":null,"position":68,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Dimensionality reduction and visualization","is_correct":true,"position":1,"images":[]},{"text":"Supervised classification only","is_correct":false,"position":2,"images":[]},{"text":"Decision tree pruning","is_correct":false,"position":3,"images":[]},{"text":"Text tokenization only","is_correct":false,"position":4,"images":[]}]},{"text":"What information does PCA use when finding components?","explanation":null,"position":69,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Class labels only","is_correct":false,"position":1,"images":[]},{"text":"Correlations and variance structure in the data","is_correct":true,"position":2,"images":[]},{"text":"Random cluster numbers","is_correct":false,"position":3,"images":[]},{"text":"Only the target variable","is_correct":false,"position":4,"images":[]}]},{"text":"Does PCA use class labels when finding principal components?","explanation":null,"position":70,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Yes, always","is_correct":false,"position":1,"images":[]},{"text":"No, PCA is an unsupervised method","is_correct":true,"position":2,"images":[]},{"text":"Only for binary classification","is_correct":false,"position":3,"images":[]},{"text":"Only for logistic regression","is_correct":false,"position":4,"images":[]}]},{"text":"Principal components in PCA are:","explanation":null,"position":71,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Original features without changes","is_correct":false,"position":1,"images":[]},{"text":"New directions formed as combinations of original features","is_correct":true,"position":2,"images":[]},{"text":"Class labels","is_correct":false,"position":3,"images":[]},{"text":"Decision tree nodes","is_correct":false,"position":4,"images":[]}]},{"text":"Why can PCA components be difficult to interpret?","explanation":null,"position":72,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They contain no numerical values","is_correct":false,"position":1,"images":[]},{"text":"They are combinations of many original features","is_correct":true,"position":2,"images":[]},{"text":"They are always class labels","is_correct":false,"position":3,"images":[]},{"text":"They are created only from test data","is_correct":false,"position":4,"images":[]}]},{"text":"Why are the first two PCA components often used?","explanation":null,"position":73,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To delete the dataset","is_correct":false,"position":1,"images":[]},{"text":"To visualize high-dimensional data in two-dimensional space","is_correct":true,"position":2,"images":[]},{"text":"To build a Random Forest","is_correct":false,"position":3,"images":[]},{"text":"To calculate K in K-Means","is_correct":false,"position":4,"images":[]}]},{"text":"In image analysis, PCA can be used to:","explanation":null,"position":74,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Extract important patterns and reduce image dimensionality","is_correct":true,"position":1,"images":[]},{"text":"Create target labels manually","is_correct":false,"position":2,"images":[]},{"text":"Ignore pixel values","is_correct":false,"position":3,"images":[]},{"text":"Replace all classifiers permanently","is_correct":false,"position":4,"images":[]}]},{"text":"What is whitening commonly related to in PCA preprocessing?","explanation":null,"position":75,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Making all images white","is_correct":false,"position":1,"images":[]},{"text":"Transforming components to have comparable scale and reduced correlation","is_correct":true,"position":2,"images":[]},{"text":"Removing the training set","is_correct":false,"position":3,"images":[]},{"text":"Changing class labels into numbers","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main purpose of NMF?","explanation":null,"position":76,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Supervised classification","is_correct":false,"position":1,"images":[]},{"text":"Feature extraction using non-negative components","is_correct":true,"position":2,"images":[]},{"text":"Creating decision trees","is_correct":false,"position":3,"images":[]},{"text":"Scaling values to negative ranges","is_correct":false,"position":4,"images":[]}]},{"text":"In NMF, coefficients are required to be:","explanation":null,"position":77,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Negative","is_correct":false,"position":1,"images":[]},{"text":"Non-negative","is_correct":true,"position":2,"images":[]},{"text":"Only binary labels","is_correct":false,"position":3,"images":[]},{"text":"Random strings","is_correct":false,"position":4,"images":[]}]},{"text":"Why can NMF be useful for image or signal data?","explanation":null,"position":78,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It can identify meaningful parts or components","is_correct":true,"position":1,"images":[]},{"text":"It requires class labels","is_correct":false,"position":2,"images":[]},{"text":"It always produces decision boundaries","is_correct":false,"position":3,"images":[]},{"text":"It removes all useful information","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main purpose of t-SNE?","explanation":null,"position":79,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Regression prediction","is_correct":false,"position":1,"images":[]},{"text":"Tree pruning","is_correct":false,"position":2,"images":[]},{"text":"Visualization of high-dimensional data in low-dimensional space","is_correct":true,"position":3,"images":[]},{"text":"Replacing feature scaling","is_correct":false,"position":4,"images":[]}]},{"text":"What information does t-SNE mainly use?","explanation":null,"position":80,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only target labels","is_correct":false,"position":1,"images":[]},{"text":"Distances or similarities between data points","is_correct":true,"position":2,"images":[]},{"text":"Only decision tree depth","is_correct":false,"position":3,"images":[]},{"text":"Only class names","is_correct":false,"position":4,"images":[]}]},{"text":"What is clustering?","explanation":null,"position":81,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Predicting a known target variable","is_correct":false,"position":1,"images":[]},{"text":"Grouping similar data points without using labels","is_correct":true,"position":2,"images":[]},{"text":"Calculating regression coefficients","is_correct":false,"position":3,"images":[]},{"text":"Training a logistic regression model","is_correct":false,"position":4,"images":[]}]},{"text":"In clustering, what does a cluster label represent?","explanation":null,"position":82,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The true supervised target","is_correct":false,"position":1,"images":[]},{"text":"The group assigned to a data point by the clustering algorithm","is_correct":true,"position":2,"images":[]},{"text":"The regression coefficient","is_correct":false,"position":3,"images":[]},{"text":"The model accuracy","is_correct":false,"position":4,"images":[]}]},{"text":"What does K-Means clustering do?","explanation":null,"position":83,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Predicts continuous values from labels","is_correct":false,"position":1,"images":[]},{"text":"Assigns points to clusters based on nearest cluster centers","is_correct":true,"position":2,"images":[]},{"text":"Builds a Naive Bayes classifier","is_correct":false,"position":3,"images":[]},{"text":"Finds class probabilities","is_correct":false,"position":4,"images":[]}]},{"text":"What does K represent in K-Means?","explanation":null,"position":84,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Number of features","is_correct":false,"position":1,"images":[]},{"text":"Number of clusters","is_correct":true,"position":2,"images":[]},{"text":"Number of test samples","is_correct":false,"position":3,"images":[]},{"text":"Number of principal components","is_correct":false,"position":4,"images":[]}]},{"text":"What are cluster centers in K-Means?","explanation":null,"position":85,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Decision tree leaves","is_correct":false,"position":1,"images":[]},{"text":"Representative points of clusters","is_correct":true,"position":2,"images":[]},{"text":"Class probabilities","is_correct":false,"position":3,"images":[]},{"text":"Regression target values","is_correct":false,"position":4,"images":[]}]},{"text":"Which is a limitation of K-Means?","explanation":null,"position":86,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It never requires choosing the number of clusters","is_correct":false,"position":1,"images":[]},{"text":"It can struggle with complex cluster shapes","is_correct":true,"position":2,"images":[]},{"text":"It cannot use numerical features","is_correct":false,"position":3,"images":[]},{"text":"It is always supervised","is_correct":false,"position":4,"images":[]}]},{"text":"Why can random initialization affect K-Means results?","explanation":null,"position":87,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"K-Means never uses initial centers","is_correct":false,"position":1,"images":[]},{"text":"Different initial centers can lead to different final clusters","is_correct":true,"position":2,"images":[]},{"text":"Initialization removes all features","is_correct":false,"position":3,"images":[]},{"text":"Initialization changes labels into probabilities","is_correct":false,"position":4,"images":[]}]},{"text":"In the context of K-Means, vector quantization means:","explanation":null,"position":88,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Representing data points using cluster centers","is_correct":true,"position":1,"images":[]},{"text":"Training a logistic regression model","is_correct":false,"position":2,"images":[]},{"text":"Calculating class probabilities","is_correct":false,"position":3,"images":[]},{"text":"Removing all clusters","is_correct":false,"position":4,"images":[]}]},{"text":"What is agglomerative clustering?","explanation":null,"position":89,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A bottom-up hierarchical clustering method","is_correct":true,"position":1,"images":[]},{"text":"A supervised classification algorithm","is_correct":false,"position":2,"images":[]},{"text":"A linear regression method","is_correct":false,"position":3,"images":[]},{"text":"A feature scaling technique","is_correct":false,"position":4,"images":[]}]},{"text":"What does hierarchical clustering produce?","explanation":null,"position":90,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A logistic regression equation","is_correct":false,"position":1,"images":[]},{"text":"A tree-like structure showing relationships between clusters","is_correct":true,"position":2,"images":[]},{"text":"Only one regression coefficient","is_correct":false,"position":3,"images":[]},{"text":"A confusion matrix only","is_correct":false,"position":4,"images":[]}]},{"text":"What is a dendrogram used for?","explanation":null,"position":91,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Visualizing the hierarchy of clusters","is_correct":true,"position":1,"images":[]},{"text":"Showing logistic regression coefficients","is_correct":false,"position":2,"images":[]},{"text":"Scaling numerical data","is_correct":false,"position":3,"images":[]},{"text":"Testing Random Forest parallelism","is_correct":false,"position":4,"images":[]}]},{"text":"What type of clustering algorithm is DBSCAN?","explanation":null,"position":92,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Linear model based clustering","is_correct":false,"position":1,"images":[]},{"text":"Density-based clustering algorithm","is_correct":true,"position":2,"images":[]},{"text":"Supervised regression algorithm","is_correct":false,"position":3,"images":[]},{"text":"Tree ensemble classifier","is_correct":false,"position":4,"images":[]}]},{"text":"What can DBSCAN identify in addition to clusters?","explanation":null,"position":93,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Regression coefficients","is_correct":false,"position":1,"images":[]},{"text":"Noise or outlier points","is_correct":true,"position":2,"images":[]},{"text":"Principal components","is_correct":false,"position":3,"images":[]},{"text":"Feature importance from trees","is_correct":false,"position":4,"images":[]}]},{"text":"In DBSCAN, what does epsilon define?","explanation":null,"position":94,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Number of trees","is_correct":false,"position":1,"images":[]},{"text":"Neighborhood radius around a point","is_correct":true,"position":2,"images":[]},{"text":"Number of classes","is_correct":false,"position":3,"images":[]},{"text":"Learning rate of logistic regression","is_correct":false,"position":4,"images":[]}]},{"text":"In DBSCAN, what does minPts define?","explanation":null,"position":95,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Maximum number of features","is_correct":false,"position":1,"images":[]},{"text":"Minimum number of points required to form a dense region","is_correct":true,"position":2,"images":[]},{"text":"Number of PCA components","is_correct":false,"position":3,"images":[]},{"text":"Number of decision tree levels","is_correct":false,"position":4,"images":[]}]},{"text":"Why is AI considered an infrastructure layer in modern information systems?","explanation":null,"position":96,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Because AI only replaces hardware","is_correct":false,"position":1,"images":[]},{"text":"Because AI can be embedded into databases, workflows, pipelines, and decision processes","is_correct":true,"position":2,"images":[]},{"text":"Because AI removes the need for governance","is_correct":false,"position":3,"images":[]},{"text":"Because AI works only as a standalone chatbot","is_correct":false,"position":4,"images":[]}]},{"text":"What is a key characteristic of agentic information systems?","explanation":null,"position":97,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They can only store files","is_correct":false,"position":1,"images":[]},{"text":"They can plan and perform multi-step tasks with some autonomy","is_correct":true,"position":2,"images":[]},{"text":"They do not interact with users or tools","is_correct":false,"position":3,"images":[]},{"text":"They are only used for static reports","is_correct":false,"position":4,"images":[]}]},{"text":"Why is human-in-the-loop control important in agentic AI systems?","explanation":null,"position":98,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To remove accountability","is_correct":false,"position":1,"images":[]},{"text":"To supervise high-risk decisions and manage unexpected behavior","is_correct":true,"position":2,"images":[]},{"text":"To prevent all automation","is_correct":false,"position":3,"images":[]},{"text":"To delete audit logs","is_correct":false,"position":4,"images":[]}]},{"text":"What is the purpose of MLOps?","explanation":null,"position":99,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To avoid deploying models","is_correct":false,"position":1,"images":[]},{"text":"To manage model training, deployment, monitoring, versioning, and validation","is_correct":true,"position":2,"images":[]},{"text":"To replace all databases with spreadsheets","is_correct":false,"position":3,"images":[]},{"text":"To remove data governance","is_correct":false,"position":4,"images":[]}]},{"text":"What is model drift?","explanation":null,"position":100,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A model becoming physically damaged","is_correct":false,"position":1,"images":[]},{"text":"A change in data or environment that reduces model performance over time","is_correct":true,"position":2,"images":[]},{"text":"A method for clustering images","is_correct":false,"position":3,"images":[]},{"text":"A type of feature scaling","is_correct":false,"position":4,"images":[]}]},{"text":"What does Cloud 3.0 emphasize?","explanation":null,"position":101,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only local storage","is_correct":false,"position":1,"images":[]},{"text":"Integrated services, real-time responsiveness, hybrid environments, and platform engineering","is_correct":true,"position":2,"images":[]},{"text":"Removing APIs from systems","is_correct":false,"position":3,"images":[]},{"text":"Using no cloud services at all","is_correct":false,"position":4,"images":[]}]},{"text":"What is a hybrid cloud environment?","explanation":null,"position":102,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A system using only personal laptops","is_correct":false,"position":1,"images":[]},{"text":"A combination of on-premises, edge, and cloud resources","is_correct":true,"position":2,"images":[]},{"text":"A model with no security controls","is_correct":false,"position":3,"images":[]},{"text":"A database without networking","is_correct":false,"position":4,"images":[]}]},{"text":"What is a data fabric intended to provide?","explanation":null,"position":103,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only manual data entry","is_correct":false,"position":1,"images":[]},{"text":"Consistent metadata, lineage, and access control across data sources","is_correct":true,"position":2,"images":[]},{"text":"Random model predictions","is_correct":false,"position":3,"images":[]},{"text":"Removal of all databases","is_correct":false,"position":4,"images":[]}]},{"text":"What is the purpose of FinOps in cloud environments?","explanation":null,"position":104,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To remove cloud monitoring","is_correct":false,"position":1,"images":[]},{"text":"To manage and optimize cloud costs through observability and accountability","is_correct":true,"position":2,"images":[]},{"text":"To train decision trees","is_correct":false,"position":3,"images":[]},{"text":"To perform PCA on images","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main goal of data governance?","explanation":null,"position":105,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To make data unavailable","is_correct":false,"position":1,"images":[]},{"text":"To ensure data quality, security, compliance, lineage, and trustworthy use","is_correct":true,"position":2,"images":[]},{"text":"To delete metadata","is_correct":false,"position":3,"images":[]},{"text":"To avoid access control","is_correct":false,"position":4,"images":[]}]},{"text":"What does data lineage describe?","explanation":null,"position":106,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The number of clusters in a dataset","is_correct":false,"position":1,"images":[]},{"text":"The origin, movement, and transformation history of data","is_correct":true,"position":2,"images":[]},{"text":"The depth of a decision tree","is_correct":false,"position":3,"images":[]},{"text":"The number of AI agents","is_correct":false,"position":4,"images":[]}]},{"text":"What is the benefit of real-time analytics?","explanation":null,"position":107,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It delays all decisions","is_correct":false,"position":1,"images":[]},{"text":"It enables decisions based on current data at the moment they matter","is_correct":true,"position":2,"images":[]},{"text":"It removes the need for data streams","is_correct":false,"position":3,"images":[]},{"text":"It only works without automation","is_correct":false,"position":4,"images":[]}]},{"text":"What is an event-driven architecture useful for?","explanation":null,"position":108,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only static yearly reporting","is_correct":false,"position":1,"images":[]},{"text":"Reacting to changes and events as they occur","is_correct":true,"position":2,"images":[]},{"text":"Removing communication between systems","is_correct":false,"position":3,"images":[]},{"text":"Avoiding data integration","is_correct":false,"position":4,"images":[]}]},{"text":"What is a feature store used for in machine learning systems?","explanation":null,"position":109,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To store only raw images","is_correct":false,"position":1,"images":[]},{"text":"To manage and serve model input features consistently","is_correct":true,"position":2,"images":[]},{"text":"To replace model evaluation","is_correct":false,"position":3,"images":[]},{"text":"To delete training datasets","is_correct":false,"position":4,"images":[]}]},{"text":"What is continuous monitoring in cybersecurity?","explanation":null,"position":110,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Checking systems once a year","is_correct":false,"position":1,"images":[]},{"text":"Ongoing observation of activity to detect suspicious behavior","is_correct":true,"position":2,"images":[]},{"text":"Removing logs after deployment","is_correct":false,"position":3,"images":[]},{"text":"Disabling identity controls","is_correct":false,"position":4,"images":[]}]},{"text":"What is the main idea of Zero Trust security?","explanation":null,"position":111,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Trust every internal user automatically","is_correct":false,"position":1,"images":[]},{"text":"Never trust implicitly and verify every access request","is_correct":true,"position":2,"images":[]},{"text":"Remove authentication","is_correct":false,"position":3,"images":[]},{"text":"Allow unlimited access to all systems","is_correct":false,"position":4,"images":[]}]},{"text":"What does the principle of least privilege mean?","explanation":null,"position":112,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Give every user administrator rights","is_correct":false,"position":1,"images":[]},{"text":"Give users and services only the access they need","is_correct":true,"position":2,"images":[]},{"text":"Disable all permissions","is_correct":false,"position":3,"images":[]},{"text":"Share one password across all systems","is_correct":false,"position":4,"images":[]}]},{"text":"How can attackers misuse AI?","explanation":null,"position":113,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"AI cannot be used by attackers","is_correct":false,"position":1,"images":[]},{"text":"For automated phishing, imitation, and model attacks","is_correct":true,"position":2,"images":[]},{"text":"Only for improving system backup","is_correct":false,"position":3,"images":[]},{"text":"Only for scaling databases","is_correct":false,"position":4,"images":[]}]},{"text":"What is model poisoning?","explanation":null,"position":114,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Improving model accuracy with clean data","is_correct":false,"position":1,"images":[]},{"text":"Manipulating training data or learning process to harm model behavior","is_correct":true,"position":2,"images":[]},{"text":"Compressing images with PCA","is_correct":false,"position":3,"images":[]},{"text":"Scaling features to zero mean","is_correct":false,"position":4,"images":[]}]},{"text":"What is a membership inference attack?","explanation":null,"position":115,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A method for increasing cluster count","is_correct":false,"position":1,"images":[]},{"text":"An attack that tries to determine whether a specific record was used in model training","is_correct":true,"position":2,"images":[]},{"text":"A decision tree pruning method","is_correct":false,"position":3,"images":[]},{"text":"A type of linear regression","is_correct":false,"position":4,"images":[]}]},{"text":"What is red teaming in AI and cybersecurity?","explanation":null,"position":116,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Deleting model documentation","is_correct":false,"position":1,"images":[]},{"text":"Testing systems by simulating adversarial attacks","is_correct":true,"position":2,"images":[]},{"text":"Training only unsupervised models","is_correct":false,"position":3,"images":[]},{"text":"Scaling data using MinMaxScaler","is_correct":false,"position":4,"images":[]}]},{"text":"What is confidential computing designed to protect?","explanation":null,"position":117,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only printed documents","is_correct":false,"position":1,"images":[]},{"text":"Data in use through secure hardware-backed environments","is_correct":true,"position":2,"images":[]},{"text":"Only public webpages","is_correct":false,"position":3,"images":[]},{"text":"Only cluster centers","is_correct":false,"position":4,"images":[]}]},{"text":"What is a domain-specific AI model?","explanation":null,"position":118,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A model trained for no particular task","is_correct":false,"position":1,"images":[]},{"text":"A model optimized for a specific industry or type of problem","is_correct":true,"position":2,"images":[]},{"text":"A model used only for random numbers","is_correct":false,"position":3,"images":[]},{"text":"A model without evaluation","is_correct":false,"position":4,"images":[]}]},{"text":"What is an AI-native architecture?","explanation":null,"position":119,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A system that avoids all models","is_correct":false,"position":1,"images":[]},{"text":"An architecture built around model runtimes, data pipelines, and continuous evaluation","is_correct":true,"position":2,"images":[]},{"text":"A paper-based workflow","is_correct":false,"position":3,"images":[]},{"text":"A single isolated spreadsheet","is_correct":false,"position":4,"images":[]}]},{"text":"What can personal AI agents do?","explanation":null,"position":120,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only turn off computers","is_correct":false,"position":1,"images":[]},{"text":"Perform multi-step personal or professional tasks such as scheduling and summarization","is_correct":true,"position":2,"images":[]},{"text":"Only create cluster labels","is_correct":false,"position":3,"images":[]},{"text":"Only run Ridge Regression","is_correct":false,"position":4,"images":[]}]},{"text":"What does Agent-as-a-Service mean?","explanation":null,"position":121,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Selling only hardware servers","is_correct":false,"position":1,"images":[]},{"text":"Providing autonomous agent capabilities as a service","is_correct":true,"position":2,"images":[]},{"text":"Removing automation from enterprise systems","is_correct":false,"position":3,"images":[]},{"text":"Using only manual workflows","is_correct":false,"position":4,"images":[]}]},{"text":"Why is AI governance important?","explanation":null,"position":122,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To prevent any model from being evaluated","is_correct":false,"position":1,"images":[]},{"text":"To ensure accountability, safety, compliance, and responsible model use","is_correct":true,"position":2,"images":[]},{"text":"To remove audit trails","is_correct":false,"position":3,"images":[]},{"text":"To hide model behavior from users","is_correct":false,"position":4,"images":[]}]},{"text":"What is explainability in AI systems?","explanation":null,"position":123,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Making models impossible to understand","is_correct":false,"position":1,"images":[]},{"text":"Providing understandable reasons or insights about model decisions","is_correct":true,"position":2,"images":[]},{"text":"Removing all model outputs","is_correct":false,"position":3,"images":[]},{"text":"Increasing only training speed","is_correct":false,"position":4,"images":[]}]},{"text":"What is a model registry used for?","explanation":null,"position":124,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Deleting old datasets","is_correct":false,"position":1,"images":[]},{"text":"Tracking model versions, metadata, and deployment status","is_correct":true,"position":2,"images":[]},{"text":"Replacing all databases","is_correct":false,"position":3,"images":[]},{"text":"Creating random forests automatically","is_correct":false,"position":4,"images":[]}]},{"text":"Why are audit trails important in AI-powered systems?","explanation":null,"position":125,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They hide system behavior","is_correct":false,"position":1,"images":[]},{"text":"They record actions and decisions for accountability and investigation","is_correct":true,"position":2,"images":[]},{"text":"They remove user permissions","is_correct":false,"position":3,"images":[]},{"text":"They replace model training","is_correct":false,"position":4,"images":[]}]},{"text":"What is technical debt?","explanation":null,"position":126,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A financial loan for buying computers","is_correct":false,"position":1,"images":[]},{"text":"Accumulated complexity or weak design choices that make future changes harder","is_correct":true,"position":2,"images":[]},{"text":"A feature scaling method","is_correct":false,"position":3,"images":[]},{"text":"A type of classification label","is_correct":false,"position":4,"images":[]}]},{"text":"What is the role of platform engineering?","explanation":null,"position":127,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To prevent developers from using tools","is_correct":false,"position":1,"images":[]},{"text":"To provide reusable internal platforms and building blocks for teams","is_correct":true,"position":2,"images":[]},{"text":"To remove all cloud services","is_correct":false,"position":3,"images":[]},{"text":"To manually label every dataset","is_correct":false,"position":4,"images":[]}]},{"text":"Why can legacy systems be a challenge for innovation?","explanation":null,"position":128,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"They always use the newest architecture","is_correct":false,"position":1,"images":[]},{"text":"They can be difficult to integrate, scale, and modernize","is_correct":true,"position":2,"images":[]},{"text":"They never contain business data","is_correct":false,"position":3,"images":[]},{"text":"They automatically support all AI models","is_correct":false,"position":4,"images":[]}]},{"text":"Why is workflow redesign important for AI adoption?","explanation":null,"position":129,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"AI gives no value in workflows","is_correct":false,"position":1,"images":[]},{"text":"Major productivity gains often require redesigning processes around automation and agents","is_correct":true,"position":2,"images":[]},{"text":"It removes the need for monitoring","is_correct":false,"position":3,"images":[]},{"text":"It prevents human oversight","is_correct":false,"position":4,"images":[]}]},{"text":"What is the ideal role of humans in many AI-powered workflows?","explanation":null,"position":130,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Humans should never supervise AI","is_correct":false,"position":1,"images":[]},{"text":"Humans supervise, audit, set policy, and handle exceptions","is_correct":true,"position":2,"images":[]},{"text":"Humans only provide random inputs","is_correct":false,"position":3,"images":[]},{"text":"Humans remove all governance rules","is_correct":false,"position":4,"images":[]}]},{"text":"How can agentic AI improve productivity?","explanation":null,"position":131,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"By slowing down all processes","is_correct":false,"position":1,"images":[]},{"text":"By automating multi-step routine workflows and reducing cycle time","is_correct":true,"position":2,"images":[]},{"text":"By removing all data","is_correct":false,"position":3,"images":[]},{"text":"By preventing tool access","is_correct":false,"position":4,"images":[]}]},{"text":"Why do AI agents need secure tool access?","explanation":null,"position":132,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To bypass all permissions","is_correct":false,"position":1,"images":[]},{"text":"To perform actions safely while respecting identity, scope, and authorization limits","is_correct":true,"position":2,"images":[]},{"text":"To delete audit logs","is_correct":false,"position":3,"images":[]},{"text":"To avoid accountability","is_correct":false,"position":4,"images":[]}]},{"text":"Why is token lifecycle management important for agents?","explanation":null,"position":133,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To make credentials permanent and unlimited","is_correct":false,"position":1,"images":[]},{"text":"To control how long and how broadly agents can access tools","is_correct":true,"position":2,"images":[]},{"text":"To remove identity checks","is_correct":false,"position":3,"images":[]},{"text":"To disable all APIs","is_correct":false,"position":4,"images":[]}]},{"text":"Why are rollback controls important in AI systems?","explanation":null,"position":134,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To prevent updates forever","is_correct":false,"position":1,"images":[]},{"text":"To return to a previous safe state when a problem occurs","is_correct":true,"position":2,"images":[]},{"text":"To delete all model versions","is_correct":false,"position":3,"images":[]},{"text":"To avoid testing","is_correct":false,"position":4,"images":[]}]},{"text":"What is continuous evaluation of AI models?","explanation":null,"position":135,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Testing a model only once before deployment","is_correct":false,"position":1,"images":[]},{"text":"Regularly checking model quality, safety, and performance over time","is_correct":true,"position":2,"images":[]},{"text":"Removing validation data","is_correct":false,"position":3,"images":[]},{"text":"Using only manual predictions","is_correct":false,"position":4,"images":[]}]},{"text":"What is bias in AI systems?","explanation":null,"position":136,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"A guaranteed improvement in accuracy","is_correct":false,"position":1,"images":[]},{"text":"Systematic unfairness or distorted behavior in model outputs","is_correct":true,"position":2,"images":[]},{"text":"A clustering center","is_correct":false,"position":3,"images":[]},{"text":"A scaling parameter","is_correct":false,"position":4,"images":[]}]},{"text":"What is privacy leakage in AI systems?","explanation":null,"position":137,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Protecting all data perfectly","is_correct":false,"position":1,"images":[]},{"text":"Unintended exposure of sensitive information","is_correct":true,"position":2,"images":[]},{"text":"Increasing model explainability only","is_correct":false,"position":3,"images":[]},{"text":"Reducing dimensionality","is_correct":false,"position":4,"images":[]}]},{"text":"Why must sensitive data be carefully controlled in AI systems?","explanation":null,"position":138,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Because it is never useful","is_correct":false,"position":1,"images":[]},{"text":"Because misuse can cause privacy, legal, and ethical risks","is_correct":true,"position":2,"images":[]},{"text":"Because it removes all model drift","is_correct":false,"position":3,"images":[]},{"text":"Because it prevents any prediction","is_correct":false,"position":4,"images":[]}]},{"text":"What does it mean to treat data as a product?","explanation":null,"position":139,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Data is ignored after collection","is_correct":false,"position":1,"images":[]},{"text":"Data is managed with quality, ownership, documentation, and service expectations","is_correct":true,"position":2,"images":[]},{"text":"Data is never reused","is_correct":false,"position":3,"images":[]},{"text":"Data is stored without metadata","is_correct":false,"position":4,"images":[]}]},{"text":"What can a data SLA define?","explanation":null,"position":140,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only the color of charts","is_correct":false,"position":1,"images":[]},{"text":"Expectations for data quality, latency, and availability","is_correct":true,"position":2,"images":[]},{"text":"The number of hidden layers in a neural network","is_correct":false,"position":3,"images":[]},{"text":"The number of clusters in K-Means","is_correct":false,"position":4,"images":[]}]},{"text":"What is data provenance?","explanation":null,"position":141,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"The final accuracy score only","is_correct":false,"position":1,"images":[]},{"text":"Information about where data came from and how it was produced","is_correct":true,"position":2,"images":[]},{"text":"The number of CPU cores used","is_correct":false,"position":3,"images":[]},{"text":"A type of Naive Bayes classifier","is_correct":false,"position":4,"images":[]}]},{"text":"What are tamper-evident logs used for?","explanation":null,"position":142,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To hide unauthorized changes","is_correct":false,"position":1,"images":[]},{"text":"To make unauthorized changes detectable","is_correct":true,"position":2,"images":[]},{"text":"To remove all monitoring","is_correct":false,"position":3,"images":[]},{"text":"To train PCA models","is_correct":false,"position":4,"images":[]}]},{"text":"What is a service mesh useful for?","explanation":null,"position":143,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Only drawing charts","is_correct":false,"position":1,"images":[]},{"text":"Managing secure communication between services","is_correct":true,"position":2,"images":[]},{"text":"Removing APIs","is_correct":false,"position":3,"images":[]},{"text":"Replacing all databases with files","is_correct":false,"position":4,"images":[]}]},{"text":"What is edge computing?","explanation":null,"position":144,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Processing data only in a distant central server","is_correct":false,"position":1,"images":[]},{"text":"Processing data closer to where it is generated","is_correct":true,"position":2,"images":[]},{"text":"Deleting local devices","is_correct":false,"position":3,"images":[]},{"text":"Using no network connection at all","is_correct":false,"position":4,"images":[]}]},{"text":"Why might latency-sensitive systems use edge or hybrid architecture?","explanation":null,"position":145,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To increase delay","is_correct":false,"position":1,"images":[]},{"text":"To reduce response time by processing closer to the source","is_correct":true,"position":2,"images":[]},{"text":"To remove all security","is_correct":false,"position":3,"images":[]},{"text":"To avoid data collection","is_correct":false,"position":4,"images":[]}]},{"text":"Why is cloud cost observability important?","explanation":null,"position":146,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"To make spending invisible","is_correct":false,"position":1,"images":[]},{"text":"To monitor and control cloud expenses continuously","is_correct":true,"position":2,"images":[]},{"text":"To remove billing responsibility","is_correct":false,"position":3,"images":[]},{"text":"To disable elastic capacity","is_correct":false,"position":4,"images":[]}]},{"text":"What is a risk of poorly governed automation?","explanation":null,"position":147,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"It always improves accountability","is_correct":false,"position":1,"images":[]},{"text":"It can cause errors at scale without proper supervision","is_correct":true,"position":2,"images":[]},{"text":"It removes all operational complexity","is_correct":false,"position":3,"images":[]},{"text":"It prevents data leakage automatically","is_correct":false,"position":4,"images":[]}]},{"text":"What is important for successful enterprise AI adoption?","explanation":null,"position":148,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Using models without monitoring","is_correct":false,"position":1,"images":[]},{"text":"Combining platforms, governance, security, data quality, and reskilling","is_correct":true,"position":2,"images":[]},{"text":"Avoiding all cloud services","is_correct":false,"position":3,"images":[]},{"text":"Ignoring model performance after deployment","is_correct":false,"position":4,"images":[]}]},{"text":"What does the idea of AI models as operating systems suggest?","explanation":null,"position":149,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Models only write documents","is_correct":false,"position":1,"images":[]},{"text":"Models may orchestrate tools, applications, data access, and workflows","is_correct":true,"position":2,"images":[]},{"text":"Models cannot interact with services","is_correct":false,"position":3,"images":[]},{"text":"Models replace all hardware","is_correct":false,"position":4,"images":[]}]},{"text":"Which practice best supports responsible AI?","explanation":null,"position":150,"is_review_required":false,"answer_text":null,"images":[],"options":[{"text":"Deploying models without testing","is_correct":false,"position":1,"images":[]},{"text":"Monitoring, explainability, governance, privacy protection, and human oversight","is_correct":true,"position":2,"images":[]},{"text":"Removing access controls","is_correct":false,"position":3,"images":[]},{"text":"Ignoring bias and drift","is_correct":false,"position":4,"images":[]}]}]}