Transfer Learning with CNNs Flashcards

Which layer type is typically removed when adapting a pretrained CNN for a new classification task?

Answer: Fully connected output layer

The final fully connected (classification) layer is replaced because it was trained for the original task's class labels.

When fine-tuning only the last few layers of a pretrained CNN, what is the main concern if you use a very high learning rate?

Answer: Catastrophic forgetting of pretrained features

A high learning rate can overwrite the valuable pretrained weights through catastrophic forgetting, destroying learned representations.

What does 'feature extraction' mode mean in transfer learning?

Answer: Freezing pretrained layers and only training the new classifier head

Feature extraction freezes the convolutional base and trains only the added classification head on the new dataset.

A pretrained CNN trained on ImageNet is being adapted for satellite imagery classification. Which transfer learning strategy is MOST appropriate?

Answer: Fine-tune all layers with a small learning rate

Satellite imagery differs significantly from ImageNet photos, so fine-tuning all layers (with a small LR) allows the model to adapt its features.

Which metric best indicates whether a source domain is suitable for transfer to a target domain?

Answer: Domain similarity or distribution overlap

Domain similarity determines how transferable the learned features are; high similarity means better transfer.

In progressive fine-tuning, layers are unfrozen from which direction?

Answer: From the last (output) layer backward

Progressive (gradual) fine-tuning unfreezes layers starting from the top (task-specific) toward the bottom (general) to preserve low-level features.

What is 'negative transfer' in the context of transfer learning with CNNs?

Answer: When pretraining hurts performance on the target task

Negative transfer occurs when features from the source domain are incompatible with the target domain, degrading target task performance.