Question 7When creating deep learning algorithms, developers configure the number of layers and the type of functions that connect the outputs of each layer to the inputs of the next.1 pointTrueFalse

How many hidden layers might you find in a deep learning system? Question 11 options: Never more than six Thirty or more At least ten Two or more

You1 pointWhich of the following is not a type of layer in a neural network? Input layer Hidden layer Output layer Support layer1 pointWhat is the process of adjusting control parameters to optimize a neural network's performance called? Regularization Hyperparameter tuning Gradient descent Feature scaling1 pointWhat is the purpose of the learning rate in a neural network? To control the speed of weight updates To determine the number of layers To set the activation function To initialize the weights1 pointWhat is the purpose of the loss function in a neural network? To measure the accuracy of the model To update the weights To compute the gradient To measure the difference between predicted output and actual output1 pointWhat does the term 'backpropagation' refer to in neural networks? Forward movement of information in a neural network Fine-tuning the weights by propagating errors backward Activation of output neurons Weight initialization process1 pointWhich algorithm is commonly used for updating weights in backpropagation? Gradient Descent K-Means Random Forest Principal Component Analysis1 pointWhat does the term 'epoch' refer to in neural network training? A type of activation function Number of layers in a network One complete cycle of training data through the network A method for weight initialization1 pointWhat is a perceptron? a single layer feed-forward neural network an auto-associative neural network a double layer auto-associative neural network a neural network that contains feedback1 pointWhich of the following best defines cross-sectional data? Data collected over different time periods from the same subjects Data collected from a single point in time from different subjects Data collected from the same subjects over multiple time points Data collected from a specific population at regular intervals1 pointIf a neural network has 16 input neurons and 4 output neurons, how many neurons would be recommended for the hidden layer according to thumb rule? 8 neurons 4 neurons 2 neurons 12 neurons1 pointIf you increase the number of hidden layers in a multi-layer perceptron, the classification error of test data always decreases True False1 pointThere is a feedback loop in the final stage of a back propagation algorithm True False1 pointIn time series analysis, which component represents the long-term movement or the general direction of the data? Seasonality Cyclical variations Trend Residual or noise1 pointWhat defines panel data in econometric studies? Data that involve repeated multi-dimensional observations of the same subjects over different periods of time same as cohort study repeated observations at same time All the above1 pointWhat differentiates a feedforward neural network from other types of neural networks like recurrent neural networks (RNNs) or convolutional neural networks (CNNs)? It incorporates feedback connections It's designed specifically for sequential data Information flows in a single direction, without loops or cycles. It employs pooling layers for feature extraction.

Question 4.Regarding CNN Deep Learning Architectures for Computer Vision, please discussand explain the following concepts.(a) The Universal Approximation Theorem states that a 2-layer network canrepresent any computational problem to an arbitrary degree of accuracy.Why then is there a need to look at deeper networks?(b) Why is there a need for non-linear activation functions between layers?(c) Why do we use a loss function for training instead of a more intuitivemetric such as accuracy?(d) With respect to image data augmentation, how does fastai presizing workand why is it advantageous?(e) What is fine tuning and how is it used? What are the advantages of finetuning over training from scratch?(10 marks)(Part B: Total 40 marks)

Question 2What sets deep learning apart from traditional neural networks?1 pointMultiple layers of neural networksEnhanced cloud computing integrationImproved computational efficiencyLinear transformations in data analysis

Suppose we have the following convolutional neural network architecture:Input volume: 15x15x1 (i.e. width = 15, height = 15, and depth = 1)First convolutional layer: 5 filters, with receptive field 3x3x1Second convolutional layer: 5 filters, with receptive field 2x2x5Output layer: fully connected with 10 neurons.How many weights are there in each layer?