Which of the following statements about transfer learning in CNNs are correct? (mark all that apply)多项选择题

A

Pre-trained CNNs can be used for transfer learning even when the source task is very different from the target task, provided sufficient labeled data is available.

B

Fine-tuning all layers of a pre-trained CNN is always better than freezing earlier layers and only training the final layer

C

Transfer learning works best when the source and target domains are similar in terms of data distribution and task objectives.

D

Transfer learning typically improves performance because pre-trained weights act as an effective form of regularization, reducing the risk of overfitting

登录即可查看完整答案

我们收录了全球超50000道真实原题与详细解析,现在登录,立即获得答案。

类似问题

A medical team wants an LLM that understands radiology terminology but cannot afford to retrain a model from scratch. Which approach is most appropriate?

Question at position 10 What is transfer learning in neural networks? Using a pre-trained neural network to extract features from a new datasetUsing a pre-trained neural network as a starting point for training a new modelUsing a pre-trained neural network to generate new dataNone of the above

Match the following to the most appropriate descriptions for each. 1: A type of machine learning where models are pre-trained usually on abundant data from one domain, and then later trained and specialized on usually a smaller dataset from another domain 2: Training a model typically on a large dataset in a given domain in order to leverage that model in other specialized domains and use cases 3: Training a pre-trained model on data in a different domain in order to create specialized models for certain tasks in the new domain 4: Layers of a deep learning architecture whose parameters are learned during the pre-training process of transfer learning

Transfer learning is invariably effective. eg. Irrespective of the amount of data, we can always rely on transfer learning.

更多留学生实用工具

加入我们，立即解锁 海量真题 与 独家解析，让复习快人一步！