Mid-levelMultiple choiceWhat is a disadvantage of deep learning compared to traditional machine learning?ADeep learning struggles with high-dimensional data.BDeep learning has a higher computational cost.CDeep learning requires less data.DDeep learning is easier to interpret.Check answer
Mid-levelMultiple choiceWhat are autoencoders and what are the different layers involved in their architecture?AAutoencoders are used for supervised learning and consist of input, hidden, and output layers.BAutoencoders are used for semi-supervised learning and consist of input, hidden, and output layers.CAutoencoders are used for unsupervised learning and consist of input, encoder, bottleneck, decoder, and output layers.DAutoencoders are used for reinforcement learning and consist of input, encoder, and output layers.Check answer
JuniorMultiple choiceWhat is an activation function and what are three types of activation functions?AAn activation function is a mathematical operation that introduces linearity into a model. Types include Linear, ReLU, and Tanh.BAn activation function is a mathematical operation that introduces non-linearity into a model. Types include Sigmoid, ReLU, and Softmax.CAn activation function is a mathematical operation that introduces non-linearity into a model. Types include Linear, ReLU, and Tanh.DAn activation function is a mathematical operation that introduces linearity into a model. Types include Sigmoid, Linear, and Softmax.Check answer
Mid-levelMultiple choiceWhat can you do to reduce overfitting in a deep neural network?ADecrease the training data size.BUse data augmentation, add dropout, and apply regularization techniques.CIncrease the model complexity by adding more layers.DIgnore validation loss and focus on training loss.Check answer
Mid-levelMultiple choiceWhy should we use Batch Normalization in deep learning models?ATo ensure the model weights always remain positive.BTo standardize inputs to a layer for each mini-batch and improve training stability.CTo reduce the number of layers in a model.DTo increase the model's complexity.Check answer
