For Beginners Flashcards

What is a 'hyperparameter' in the context of machine learning?

Answer: A configuration set before training that controls the learning process

Hyperparameters like learning rate and layer count are set by the engineer before training, not learned from data.

Which technique randomly drops neurons during training to reduce overfitting?

Answer: Dropout

Dropout randomly deactivates a fraction of neurons during each training step, forcing the network to learn redundant representations.

What does 'transfer learning' allow an AI engineer to do?

Answer: Reuse a pre-trained model's knowledge for a new but related task

Transfer learning leverages representations learned on a large dataset (e.g., ImageNet) to accelerate training on a smaller target task.

In a confusion matrix for binary classification, what does a 'false positive' represent?

Answer: The model predicted positive but the true label is negative

A false positive (Type I error) occurs when the model incorrectly labels a negative example as positive.

What is the role of the learning rate in gradient descent?

Answer: It controls how large a step is taken in the direction of the gradient

The learning rate scales the gradient update; too high causes divergence, too low causes very slow convergence.

Which of the following best describes an embedding in AI?

Answer: A dense vector representation of data in a continuous space

Embeddings map discrete objects (words, items) to dense vectors so that similar objects are close together in vector space.

What is 'data augmentation' primarily used for in AI training?

Answer: Artificially expanding the training set by applying transformations to existing samples

Data augmentation applies transforms like flipping, cropping, or noise addition to create more varied training examples from existing data.