National Highways & Infrastructure Development Corporation\ Interview Questions and Answers

To rotate an image in Python, use the Pillow library's rotate() method.

• Import the Image module from the Pillow library

• Open the image using the open() method

• Use the rotate() method to rotate the image by the desired angle

• Save the rotated image using the save() method

CNN is used for image recognition while RNN is used for sequence data like text or speech.

• CNN is Convolutional Neural Network, used for image recognition tasks.

• RNN is Recurrent Neural Network, used for sequence data like text or speech.

• CNN has convolutional layers for feature extraction, while RNN has recurrent connections for sequential data processing.

• CNN is good at capturing spatial dependencies in data, while RNN is good at capturing temporal dependencies.

• Example: CNN can...read more

Activation functions are used to introduce non-linearity into neural networks, allowing them to learn complex patterns and relationships.

• Activation functions help neural networks to learn complex patterns and relationships by introducing non-linearity.

• They help in controlling the output of a neuron, ensuring that it falls within a desired range.

• Common activation functions include ReLU, Sigmoid, Tanh, and Leaky ReLU.

• Without activation functions, neural networks would simply be...read more

Yes, Logloss function is differentiable.

• Logloss function is differentiable as it is a smooth and continuous function.

• The derivative of Logloss function can be calculated using calculus.

• Differentiability is important for optimization algorithms like gradient descent to converge smoothly.

• Example: The derivative of Logloss function for binary classification is (predicted probability - actual label).

Learning rate is a hyperparameter that controls how much we are adjusting the weights of our network with respect to the loss gradient.

• Learning rate determines the size of the steps taken during optimization.

• A high learning rate can cause the model to overshoot the optimal weights, while a low learning rate can result in slow convergence.

• Common learning rate values are 0.1, 0.01, 0.001, etc.

• Learning rate can be adjusted during training using techniques like learning rate sche...read more