Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

Chelsea Finn, Pieter Abbeel, Sergey Levine
3/9/2017
cs.LGcs.AIcs.CV

Official code: cbfinn/maml (2,725 stars, MIT, last commit 2020-01-19). 1 more implementation below.

Abstract

We propose an algorithm for meta-learning that is model-agnostic, in the sense that it is compatible with any model trained with gradient descent and applicable to a variety of different learning problems, including classification, regression, and reinforcement learning. The goal of meta-learning is to train a model on a variety of learning tasks, such that it can solve new learning tasks using only a small number of training samples. In our approach, the parameters of the model are explicitly trained such that a small number of gradient steps with a small amount of training data from a new task will produce good generalization performance on that task. In effect, our method trains the model to be easy to fine-tune. We demonstrate that this approach leads to state-of-the-art performance on two few-shot image classification benchmarks, produces good results on few-shot regression, and accelerates fine-tuning for policy gradient reinforcement learning with neural network policies.

View on arXivView PDF

Code Implementations(2)

cbfinn/mamlOfficial100%
MIT

Code for "Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks"

2,725622PythonJun 18, 20176 years agoMIT
cbfinn/maml_rlOfficial100%

Code for RL experiments in "Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks"

672189CSS, HTMLJul 19, 20173 years agoNOASSERTION

Similar Papers

Cooperative Meta-Learning with Gradient Augmentation

91% match1 repo62024

Improving Generalization of Meta-Learning with Inverted Regularization at Inner-Level

86% match2026

Rethinking Meta-Learning from a Learning Lens

86% match2024

Task-Specific Preconditioner for Cross-Domain Few-Shot Learning

85% match2024

On the Stability and Generalization of Meta-Learning

85% match2026

Meta-Learning and representation learner: A short theoretical note

85% match2024

Cite this paper

@article{finn2017modelagnostic,
  title  = {Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks},
  author = {Chelsea Finn and Pieter Abbeel and Sergey Levine},
  year   = {2017},
  eprint = {1703.03400},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1703.03400}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1703.03400.svg)](https://codeofpaper.com/paper/1703.03400)

Discussion