Data Science & AI Interview Questions — Flashcards | inGrade

Question 1

What is Deep Learning?

Accepted Answer

Deep Learning is a branch of Machine Learning based on artificial neural networks, particularly deep neural networks. It involves layers of interconnected neurons that process data hierarchically, learning increasingly abstract representations. Deep Learning is widely used for tasks like image recognition, speech processing, and natural language understanding.

Question 2

How is Deep Learning different from Machine Learning?

Accepted Answer

Machine Learning involves traditional algorithms like decision trees or linear regression, which often rely on feature engineering. In contrast, Deep Learning automates feature extraction using multiple network layers, enabling it to handle complex and unstructured data such as images or text.

Question 3

What are the key components of a neural network?

Accepted Answer

A neural network consists of: Input Layer (receives data inputs), Hidden Layers (process data through weights and activation functions), Output Layer (provides predictions), Weights and Biases (adjust based on learning), and Activation Functions (introduce non-linearity for complex learning).

Question 4

What role does backpropagation play in Deep Learning?

Accepted Answer

Backpropagation adjusts weights and biases to minimize error by calculating gradients using the chain rule. It is critical for training neural networks by iteratively improving their predictions.

Question 5

What is a Convolutional Neural Network (CNN)?

Accepted Answer

CNNs are designed for spatial data like images. They use convolutional layers to extract features and pooling layers for dimensionality reduction. They are widely used in image recognition, object detection, facial recognition, and medical image analysis.

Question 6

What is Transfer Learning in Deep Learning?

Accepted Answer

Transfer Learning involves leveraging pre-trained models for a new, related task. For example, using a model trained on ImageNet to classify medical images. It reduces training time, requires less data, and often improves performance by utilizing knowledge from related tasks.

Question 7

What is overfitting in Deep Learning, and how can it be prevented?

Accepted Answer

Overfitting occurs when a model learns patterns from the training data, including noise, to the extent that it performs poorly on unseen data. Prevention methods include using more training data, applying regularization (L1, L2), incorporating dropout layers, early stopping, and simplifying the model architecture.

Question 8

What is a Generative Adversarial Network (GAN)?

Accepted Answer

A GAN consists of two networks: a generator that creates synthetic data and a discriminator that distinguishes real data from generated data. They compete in a minimax game, with the generator improving to fool the discriminator. Applications include image synthesis, style transfer, and data augmentation.

Question 9

What is the vanishing gradient problem?

Accepted Answer

In deep networks, gradients diminish as they backpropagate through layers, slowing learning. This makes it difficult for the network to learn long-range dependencies. Techniques like ReLU activation functions and batch normalization help mitigate this problem.

Question 10

What are Transformers, and why are they important?

Accepted Answer

Transformers use attention mechanisms to process sequences in parallel, unlike RNNs. They power modern NLP models like GPT and BERT. They are efficient due to parallelization, scalability with longer sequences, and the ability to capture global context through self-attention.

Start your learning journey today