Fundamentals of Deep Learning

What is a neuron in the context of neural networks?

• A neuron is a biological cell that does not interact with artificial intelligence.

• A neuron is a basic processing unit in neural networks that receives inputs, applies weights, and produces an output.

• A neuron is a type of computer hardware used for processing data.

• A neuron is a software algorithm that generates random numbers.

Explain the structure of a simple neural network.

• A simple neural network has an input layer, hidden layers, and an output layer, with neurons connected by adjustable weights.

• A simple neural network consists of only an input layer and an output layer without any hidden layers.

• A simple neural network has multiple output layers and no input layer.

• Neurons in a simple neural network are not connected by weights but by fixed values.

What role do activation functions play in neural networks?

• Activation functions enable neural networks to learn complex patterns by introducing non-linearity.

• Activation functions only work in the output layer of a neural network.

• Activation functions are used to increase the speed of training.

• Activation functions are responsible for data normalization.

Define overfitting and how it can be mitigated.

• Overfitting is when a model performs poorly on both training and unseen data.

• Overfitting occurs when a model is too simple and cannot learn from the data.

• Overfitting can be improved by increasing the size of the training dataset only.

• Overfitting is when a model performs well on training data but poorly on unseen data. It can be mitigated by using techniques like regularization, cross-validation, and simplifying the model.

What is the purpose of the loss function in training a neural network?

• To select the training dataset for the model.

• To optimize the learning rate of the model.

• To measure the error between predicted and actual values, guiding weight adjustments during training.

• To determine the architecture of the neural network.

Describe the concept of gradient descent.

• Gradient descent is an optimization algorithm that minimizes a function by iteratively moving in the direction of the negative gradient.

• Gradient descent is a technique used to visualize data in three dimensions.

• Gradient descent is an algorithm that sorts data in ascending order.

• Gradient descent is a method for increasing a function's value by following the positive gradient.

What is the difference between batch gradient descent and stochastic gradient descent?

• Stochastic gradient descent requires more memory than batch gradient descent.

• Batch gradient descent uses the whole dataset for each update, while stochastic gradient descent uses one data point at a time.

• Batch gradient descent is faster than stochastic gradient descent for large datasets.

• Batch gradient descent updates weights more frequently than stochastic gradient descent.

Explain the term 'learning rate' in gradient descent.

• The learning rate is a hyperparameter that controls the step size in gradient descent.

• The learning rate is a measure of the model's accuracy.

• The learning rate is a fixed value that does not change during training.

• The learning rate determines the number of iterations in gradient descent.

What are convolutional neural networks (CNNs) primarily used for?

• Image recognition and classification

• Time series forecasting

• Reinforcement learning

• Natural language processing

Describe the architecture of a typical CNN.

• Input layer only

• A typical CNN architecture includes an input layer, convolutional layers, activation functions, pooling layers, fully connected layers, and an output layer.

• Convolutional layers without activation functions

• Fully connected layers without pooling layers