AWS Certified AI Practitioner (AIF-C01) — Question 249

A company has trained a foundation model (FM) to perform a specific task. The company needs to fine-tune the FM to perform a different but related task.

Which fine-tuning method will meet this requirement?

Answer options

• A. Hyperparameter tuning
• B. Pre-training
• C. Transfer learning
• D. Reinforcement learning

Correct answer: C

Explanation

The correct answer is C, Transfer learning, as it allows the foundation model to leverage knowledge from the original task and apply it to a new but related task. Hyperparameter tuning (A) adjusts the model's parameters rather than its task, Pre-training (B) involves training from scratch on a large dataset, and Reinforcement learning (D) focuses on learning through interactions rather than adapting to new tasks.