Which term describes the iterative procedure that updates model parameters by computing gradients and minimizing loss?

• A. Training Loop
• B. Prompt Engineering
• C. Generator in GAN
• D. Cloud Platforms

Answer: (A)

The training loop is the iterative process in which a model’s parameters are updated to minimize loss. In each iteration, you feed a batch of data through the model to obtain predictions, compute the loss that measures how far those predictions are from the true values, and then calculate gradients of that loss with respect to the model parameters via backpropagation. An optimizer (like SGD or Adam) uses those gradients to adjust the parameters, nudging the model toward better performance. This cycle repeats across many batches and through multiple epochs, gradually reducing the loss and improving accuracy.

Prompt engineering focuses on crafting prompts to elicit desired responses without changing model parameters, a GAN’s generator is a component that learns to produce data, and cloud platforms provide infrastructure.