Round - 3 ( Technical round )

We can categorise the BBN Butterfly, Cosmic Cube, Cedar, and Hypercube machine as

• a. MISD

• b. SIMD

• c. MIMD

• d. SISD

What is overfitting in the context of machine learning models?

• a) Fitting a model with insufficient data

• b)Fitting a model too closely to the training data

• c)Fitting a model with too few features

• d) Fitting a model to the validation set

What is feature importance in the context of tree-based models like Random Forests?

• a)The importance of including a feature in the dataset

• b)The correlation between features

• c)The contribution of each feature to the model's predictions

• d)The number of times a feature appears in the dataset

The function that assigns the symbol a to b in LISP is:

• a. (setq b ‘a’)

• b. (setq b = ‘a’)

• c. (setq b a)

• d. (set b = ‘a’)

In a neural network, what is the role of the activation function?

• a)Defines the learning rate

• b)Sets the number of neurons in each layer

• c)Introduces non-linearity to the model

• d)Controls the initialization of weights

What is the difference between bagging and boosting in ensemble learning?

• a)Bagging increases model diversity, boosting decreases it

• b)Bagging trains models sequentially, boosting trains them in parallel

• c)Bagging combines predictions using voting, boosting combines predictions using weighted averaging

• d)Bagging trains each model independently, boosting focuses on examples misclassified by previous models

What is the purpose of the Expectation-Maximization (EM) algorithm in unsupervised learning?

• a)Maximizing the likelihood of the observed data

• b)Minimizing the reconstruction error in autoencoders

• c)Imputing missing values in a dataset

• d)Iteratively estimating parameters for mixture models

What is the purpose of the epochs parameter in neural network training?

• a)The number of layers in the neural network

• b)The number of training examples processed in one iteration

• c)The learning rate for weight updates

• d)The number of complete passes through the entire training dataset

How does the choice of a kernel impact the performance of a Support Vector Machine (SVM)?

• a)The kernel has no impact on SVM performance

• b)The kernel determines the SVM's maximum margin

• c)The kernel defines the transformation of input features into a higher-dimensional space

• d)The kernel influences the learning rate during training

What is the purpose of dropout in neural networks?

• a)Reduces the model's learning rate

• b)Regularizes the model to prevent overfitting

• c)Increases the number of neurons in each layer

• d)Ensures convergence during training