Question 1 Which of the following is a potential drawback of using neural networks? O a) They are computationally efficient for all tasks. O b) They often require a large amount of labeled training data. O c) They are not suitable for parallel processing O d) They have a fixed architecture and cannot adapt to changing environments.Question 10 Which algorithm is commonly used to update the weights of a neural network during backpropagation? O a) Support Vector Machine (SVM) O b) K-Means O c) K-Nearest Neighbors (KNN) O d) Gradient DescentQuestion 11 In a feedforward neural network, information flows in which direction(S)? O a) Forward only O b) Backward only O c) Both forward and backward O d) In a random directionQuestion 12 Which type of neural network is well-suited for sequence data, such as natural language processing tasks? O a) Convolutional Neural Network (CNN) O b) Autoencoder O c) Recurrent Neural Network (RNN) O d) Deep Belief Network (DBN)Question 13 Which neural network architecture is inspired by the structure and function of the human brain's neocortex? O a) Convolutional Neural Network (CNN) O b) Recurrent Neural Network (RNN) O c) Long Short-Term Memory (LSTM) Network O d) Spiking Neural Network (SNN)Question 14 Which step is typically used to evaluate the performance of a neural network during training? O a) Testing on the training data O b) Testing on the validation data O c) Testing on the testing data O d) Testing on the entire datasetQuestion 15 Which method is used to prevent the weights of a neural network from becoming too large during training? O a) Dropout O b) L1 Regularization O c) L2 Regularization O d) Gradient ClippingQuestion 16 Which layer in a neural network is responsible for reducing the spatial dimensions of the input data in a CNN? O a) Activation layer O b) Convolutional layer O c) Fully connected layer O d) Pooling layerQuestion 17 The process of adjusting the weights of a neural network during training is called: O a) Backpropagation O b) Gradient Boosting O c) Forward Propagation O d) Cross-ValidationQuestion 18 What is the purpose of the "bias" term in a neuron? O a) It helps to regulate the learning rate of the network. O b) It prevents overfitting in the model. O c) It allows the neuron to account for a non-zero intercept in the data. O d) It helps in regularizing the weights.Question 19 What is dropout in the context of neural networks? O a) A regularization technique that randomly drops neurons during training O b) A technique to increase the learning rate during backpropagation O c) A method for reducing the number of layers in a deep neural network O d) A technique for handling missing data in the training setQuestion 2 What is the vanishing gradient problem in neural networks? O a) The model's gradients become too large, leading to instability. O b) The model gets stuck in a local minimum during training. O c) The gradients in early layers become very small, leading to slow learning. O d) The model becomes unable to generalize to new data.Question 20 Which optimization algorithm is known for its ability to accelerate the training of neural networks by adapting the learning rates of individual parameters? O a) Adam O b) Gradient Descent O c) Stochastic Gradient Descent (SGD) O d) Levenberg-MarquardtQuestion 21 In a neural network, what does the term "batch size" refer to? O a) The number of training examples processed before updating the weights O b) The number of neurons in each hidden layer O c) The number of layers in the neural network O d) The number of epochs required for convergenceQuestion 22 Which of the following is NOT a common activation function in neural networks? O a) ReLU (Rectified Linear Unit) O b) Sigmoid O c) Tanh (Hyperbolic Tangent) O d) Gradient DescentQuestion 23 The term "epoch" in neural network training refers to: O a) The number of layers in the neural network O b) The number of neurons in the input layer O c) One complete pass through the entire training dataset O d) The number of iterations required for convergenceQuestion 24 Overfitting in a neural network occurs when: a) The model is too simple to represent the data. O b) The model performs well on unseen data. O c) The model memorizes the training data and fails to generalize. O d) The model has too few parameters.Question 25 Which of the following is a hyperparameter of a neural network that controls the complexity of the model? O a) Activation function O b) Learning rate O c) Number of epochs O d) Number of training examplesQuestion 3 Which loss function is commonly used for binary classification tasks in neural networks? O a. Mean Squared Error (MSE) O b. Mean Absolute Error (MAE) O c. Cross-Entropy Loss O d. Hinge LossQuestion 4 A neural network with multiple hidden layers is referred to as: O a) Shallow neural network O b) Deep neural network O c) Wide neural network O d) Complex neural networkQuestion 5 Which neural network architecture is designed to learn to approximate a given function without requiring labeled output data? O a) Supervised Neural Network O b) Unsupervised Neural Network c) Reinforcement Learning Network O d) Semi-Supervised Neural Networkuestion 6 Which type of neural network architecture is specifically designed for image recognition tasks? O a) Recurrent Neural Network (RNN) O b) Convolutional Neural Network (CNN) O c) Generative Adversarial Network (GAN) O d) AutoencoderQuestion 7 What is a neural network? O a) A computer hardware device O b) A computational model inspired by the human brain O c) A type of operating system O d) A collection of interconnected databasesQuestion 8 What is the purpose of the softmax activation function in the output layer of a neural network? O a) It is used to avoid the vanishing gradient problem. O b) It normalizes the output probabilities, making them sum to 1. O c) It allows the network to approximate any continuous function. O d) It speeds up the convergence of the backpropagation algorithm.Question 9 Which technique can be used to visualize the learned features in the intermediate layers of a deep neural network? O a) Activation Maximization O b) Reinforcement Learning O c) Feature Scaling O d) Principal Component Analysis (PCA)