S04 - Deep Learning

Which ones are the two commonly-used techniques in unsupervised learning?

• regression

• classification

• clustering

• dimensionality reduction

Maximizing the notion of cumulative reward is the goal of which learning?

• supervised learning

• reinforcement learning

• unsupervised learning

• feature engineering

Identifying the bounding box of a face in an image is an example of ...

• multilabel classification

• scalar regression

• multiclass classification

• vector regression

The type of parameters that generally change the architecture of a model is called ...

• batch

• parameter

• hyperparameter

• weight

The following strategies are holdout strategies that can be used in evaluating a model, except ...

• k-fold validation

• simple hold out validation

• iterated k-fold validation with shuffling

• stratified sampling

Select the commonly-used data preprocessing steps applied in Machine Learning!

• model training

• feature extraction

• vectorization

• value normalization

Replacing the missing data with estimated values using the existing variables is a handling-missing-value technique called ...

• regression imputation

• multiple imputation

• pairwise deletion

• listwise deletion

Overfitting occurs when our model fails to learn any patterns from our training data.

• True

• False

These are some techniques to avoid the algorithm overfitting, except ...

• applying dropout

• applying weight regularizer

• increasing network size

• getting more data

Regularization provides constraints on the network by penalizing the model when the weights of the model are larger.

• True

• False