Glossary term
Glossary term
Training and Fine-Tuning
Transfer learning is when an AI takes what it learned from one task and uses it for another. It is like reusing knowledge, saving time, effort, and making the model smarter, faster.
Transferring information from one machine learning task to another. For example, in multi-task learning, a single model solves multiple tasks, such as a deep model that has different output nodes for different tasks. Transfer learning might involve transferring knowledge from the solution of a simpler task to a more complex one, or involve transferring knowledge from a task where there is more data to one where there is less data.
Most machine learning systems solve a single task. Transfer learning is a baby step towards artificial intelligence in which a single program can solve multiple tasks.
Fine-tuning BERT or RoBERTa for sentiment analysis is a textbook example of transfer learning.
ResNet pretrained on ImageNet is widely transferred to medical imaging and industrial inspection.
LoRA fine-tuning of Llama 3 and Mistral are modern parameter-efficient transfer learning approaches.
Created for this library
A medical imaging startup uses transfer learning from a general-purpose vision backbone to bootstrap its rare-disease classifier on a small clinical dataset.
A speech recognition vendor uses transfer learning from a general acoustic model to a client-specific call-center model.
A retail brand uses transfer learning from a foundation model to a brand-specific assistant by fine-tuning on the product catalog.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License