Glossary term
Glossary term
Training and Fine-Tuning
A subset of machine learning that discovers or improves a learning algorithm. A meta-learning system can also aim to train a model to quickly learn a new task from a small amount of data or from experience gained in previous tasks. Meta-learning algorithms generally try to achieve the following:
Improve or learn hand-engineered features (such as an initializer or an optimizer).
Be more data-efficient and compute-efficient.
Improve generalization.
Meta-learning is related to few-shot learning.
Created for this library
A medical AI team uses meta-learning to enable fast adaptation of its classifier to new rare diseases with only a few labeled examples.
A robotics research team uses meta-learning so policies trained in simulation adapt quickly to new objects in the physical world.
An NLP team uses meta-learning to bootstrap classifiers for new product categories from a few labeled customer-feedback examples.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License