Google Professional Machine Learning Engineer Question 274
Select 3Google Cloud PlatformYou are tasked with fine-tuning a foundational language model using Vertex AI Model Garden for a customer support chatbot. The goal is to adapt the model to recognize domain-specific terminology and provide accurate responses for your company’s product-related inquiries. Which steps should you take to achieve this using Vertex AI?
- A
Prepare a labeled dataset containing domain-specific examples relevant to customer support queries.
- B
Directly deploy the foundational model without any modifications to the production environment.
- C
Use Vertex AI Model Garden to select a pre-trained model that aligns with your chatbot's requirements.
- D
Apply transfer learning techniques to fine-tune the pre-trained model using your prepared dataset.
- E
Train a new model from scratch using TensorFlow or PyTorch instead of leveraging a pre-trained model.
Show answer and explanation
Correct answers: A, C, D
Explanation
Fine-tuning a foundational model using Vertex AI involves leveraging pre-trained models available in Model Garden, preparing a domain-specific dataset, and applying transfer learning techniques. These steps ensure the model is adapted to your specific use case while saving time and computational resources compared to training a new model from scratch.
- A. Correct.
Preparing a labeled dataset with domain-specific examples is essential for fine-tuning the model to understand the specific needs of your chatbot.
- B. Incorrect.
Deploying the foundational model without modifications would not allow it to recognize your domain-specific terminology or provide accurate responses.
- C. Correct.
Selecting a pre-trained model from Vertex AI Model Garden is a crucial first step because it allows you to leverage an existing high-quality model for your needs.
- D. Correct.
Transfer learning is the recommended method for fine-tuning foundational models, as it adapts the pre-trained model to your specific use case using your dataset.
- E. Incorrect.
Training a new model from scratch is unnecessary and inefficient when trying to fine-tune a foundational model, as pre-trained models already provide a solid starting point.