Transfer Learning Concepts Quiz

1. Which approach freezes all pre-trained weights and only trains new layers?

Fine-tuning

Feature extraction

Domain randomization

Multi-task learning

Submit

3. What is 'negative transfer' in the context of transfer learning?

When transfer learning hurts performance compared to training from scratch

Using negative numbers in weights

Training on negative examples only

Removing bias terms

4. In computer vision, which layer of a CNN typically captures task-specific features?

Early convolutional layers

Middle convolutional layers

Later fully connected layers

Input layer

5. True or False: Transfer learning requires the source task to be more complex than the target task.

True

False

Submit

7. Which of the following is NOT a common transfer learning strategy?

Fine-tuning all layers

Feature extraction

Random weight initialization

Domain adaptation

8. True or False: Transfer learning is less effective when you have large amounts of labeled target data.

True

False

9. What is the main goal of domain adaptation techniques?

Make source and target data distributions more similar

Increase model complexity

Remove all training data

Eliminate validation sets

Submit

11. What is transfer learning in machine learning?

Using knowledge from one task to improve performance on a different task

Training multiple models on the same dataset simultaneously

Moving data between different computing devices

Transferring weights from validation to test sets

12. True or False: BERT is a pre-trained model commonly used for NLP transfer learning tasks.

True

False

13. What is a key challenge in transfer learning?

Negative transfer when source and target are too different

Models always perform worse

Requires infinite computational power

Cannot be used with neural networks

Submit

15. Which pre-trained model is commonly used for image classification transfer learning?

ImageNet

BERT

GPT-2

Word2Vec

16. True or False: Transfer learning is only useful when the source and target tasks are identical.

True

False

Submit

18. What does domain adaptation address in transfer learning?

The difference between source and target data distributions

The number of GPUs available

The programming language used

The size of the training batch

19. In transfer learning, what is 'fine-tuning'?

Adjusting hyperparameters randomly

Unfreezing and retraining later layers of a pre-trained model

Removing all layers except the first

Increasing batch size during training

20. Which of the following is a primary advantage of transfer learning?

Eliminates the need for labeled data

Reduces training time and computational resources

Guarantees 100% accuracy on new tasks

Removes the need for model evaluation

DataAI Transfer Learning Concepts Quiz

1. Which approach freezes all pre-trained weights and only trains new layers?

2.

What first name or nickname would you like us to use?

2. A model pre-trained on a large dataset is called a ____ model.

3. What is 'negative transfer' in the context of transfer learning?

4. In computer vision, which layer of a CNN typically captures task-specific features?

5. True or False: Transfer learning requires the source task to be more complex than the target task.

6. The process of retraining a pre-trained model on a new dataset is called ____.

7. Which of the following is NOT a common transfer learning strategy?

8. True or False: Transfer learning is less effective when you have large amounts of labeled target data.

9. What is the main goal of domain adaptation techniques?

10. In medical imaging, using a model pre-trained on ImageNet for disease detection is an example of ____.

11. What is transfer learning in machine learning?

12. True or False: BERT is a pre-trained model commonly used for NLP transfer learning tasks.

13. What is a key challenge in transfer learning?

14. In transfer learning, the 'source task' refers to ____.

15. Which pre-trained model is commonly used for image classification transfer learning?

16. True or False: Transfer learning is only useful when the source and target tasks are identical.

17. Feature extraction in transfer learning involves ____.

18. What does domain adaptation address in transfer learning?

19. In transfer learning, what is 'fine-tuning'?

20. Which of the following is a primary advantage of transfer learning?