Reptile: A scalable meta-learning algorithm
📰 OpenAI News
Reptile is a scalable meta-learning algorithm that updates initial parameters towards final parameters learned on a task
Action Steps
- Sample a task
- Perform stochastic gradient descent on the task
- Update the initial parameters towards the final parameters learned on the task
- Repeat the process for multiple tasks
Who Needs to Know This
Machine learning researchers and engineers on a team can benefit from Reptile as it provides a simple and efficient way to perform meta-learning, allowing them to improve model performance on a variety of tasks
Key Insight
💡 Reptile is a scalable meta-learning algorithm that can be used with any optimizer, such as SGD
Share This
🚀 Reptile: a simple & scalable meta-learning algorithm
Key Takeaways
Reptile is a scalable meta-learning algorithm that updates initial parameters towards final parameters learned on a task
Full Article
We’ve developed a simple meta-learning algorithm called Reptile which works by repeatedly sampling a task, performing stochastic gradient descent on it, and updating the initial parameters towards the final parameters learned on that task. Reptile is the application of the Shortest Descent algorithm to the meta-learning setting, and is mathematically similar to first-order MAML (which is a version of the well-known MAML algorithm) that only needs black-box access to an optimizer such as SGD or A
DeepCamp AI