Training vs Inference - Quiz

Question 1

What is the main purpose of training an AI model?

To learn useful patterns from data and adjust model parameters To permanently store every user question To eliminate the need for data To make the model independent of all computing resources

Question 2

What is inference?

Using a trained model to process new input and produce an output Deleting the model parameters Collecting the original training dataset Training a model from scratch every time it receives input

Question 3

What normally happens to model parameters during inference?

They are generally used without being permanently updated after every input They are always completely replaced They are deleted after every prediction They are automatically doubled after every question

Question 4

What happens when you normally ask a question to a trained AI assistant?

The trained model generally uses its existing parameters to generate a response The entire model is retrained from scratch The training dataset is automatically recreated All model parameters are permanently changed by the question

Question 5

What is fine-tuning?

Additional training that adapts a trained model to a specialized task or dataset A method for deleting unwanted model parameters A type of inference that never uses data A way to increase the physical size of a computer

Question 6

Why might an AI application use an external retrieval system?

To provide information from external sources during inference To permanently replace the neural network To eliminate the need for training To guarantee that every generated answer is correct

Question 7

Why is generalization important?

It allows the model to perform usefully on new inputs It means the model memorizes every training example It means the model never needs evaluation It prevents the model from receiving new input

Question 8

Which statement best describes training and inference?

Training learns patterns and adjusts parameters; inference uses the learned model on new input Training and inference are exactly the same process Inference always retrains the model Training only happens after the model is deployed