Deep Learning Interview Questions and Answers
Question - 61 : - What do you understand by end-to-end learning?
Answer - 61 : -
It is a deep learning process where a model gets raw data as the input and all the various parts are trained simultaneously to produce the desired outcome with no intermediate tasks. The advantage of end-to-end learning is that there is no need for implicitly doing feature engineering which usually leads to a lower bias. A good example that you can quote in the content of end-to-end learning is driverless cars. They use human-provided input as guidance and are trained to automatically learn and process the information using a CNN to complete tasks.
Question - 62 : - RMSProp and Adam optimizer adjust gradients? Does this mean that they perform gradient clipping?
Answer - 62 : -
This does not inherently mean that they perform gradient clipping because gradient clipping involves setting up predetermined values beyond which the gradients cannot go, unlike Adam and RMSProp that make multiplicative adjustments to gradients.
Question - 63 : - Can you name a few hyperparameters used for training a neural network.
Answer - 63 : -
When training any neural networks there are two types of hyperparameters-one that define the structure of the neural network and the other determining how a neural network is trained. Listed are a few hyperparameters that are set before training any neural network –
- Initialization of weights
- Setting the number of hidden layers
- Learning Rate
- Number of epochs
- Activation Functions
- Batch Size
- Momentum
Question - 64 : - When is multi-task learning usually preferred?
Answer - 64 : -
Multi-task learning with deep neural networks is a subfield wherein several tasks are learned by a shared model. This reduces overfitting, enhances data efficiency, and speeds up the learning process with the use of auxiliary information. Multi-task learning is useful when there is a small amount of data for any given task and we can benefit from training a deep learning model on a large dataset.
Question - 65 : - Explain the Adam Optimizer in one minute.
Answer - 65 : -
Adaptive momentum or Adam optimizer is an optimization algorithm designed to deal with sparse gradients on noisy problems. Adam optimizer improves convergence through momentum that ensures that a model does not get stuck in saddle point and also provides per-parameter updates for faster convergence.
Question - 66 : - To what kind of problems can the cross-entropy loss function be applied?
Answer - 66 : -
- Binary Classification Problems
- Multi-Label Classification Problems
- Multi-Category Classification Problems
Question - 67 : - List the steps to implement a gradient descent algorithm.
Answer - 67 : -
- The first step is to initialize random weight and bias.
- Get values from the output layer by passing the input through the neural network.
- Determine the error between the actual and predicted value.
- Based on the neurons that contribute to the error, modify the values to minimize the error.
- Repeat the process until the optimal weights are found for the neural network.
Question - 68 : - How important is it to shuffle the training data when using batch gradient descent?
Answer - 68 : -
Shuffling the training dataset will not make much of a difference because the gradient is calculated at every epoch using the complete training dataset.
Question - 69 : - What is the benefit of using max-pooling in classification convolutional neural networks?
Answer - 69 : -
The feature maps become smaller after max-pooling in CNN and hence help reduce the computation and also give more translation in-variance. Also, we don’t lose much semantic information because we’re taking the maximum activation.
Question - 70 : - Can you name a few data structures that are commonly used in deep learning?
Answer - 70 : -
You can talk about computational graphs, tensors, matrices, data frames, and lists.