Papperoni

2017

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

Chelsea Finn, P. Abbeel, Sergey Levine

Open PDF Google Scholar

citations

Cite Score

90

AI summary

This paper introduces Model-Agnostic Meta-Learning (MAML), a novel meta-learning algorithm for fast adaptation of deep networks across classification, regression, and reinforcement learning tasks, achieving state-of-the-art few-shot image classification results and accelerating policy gradient reinforcement learning.

Main Contributions

Proposes a model- and task-agnostic meta-learning algorithm (MAML) that trains model parameters for fast adaptation to new tasks.
Demonstrates the algorithm's effectiveness across different model types (fully connected, convolutional networks) and domains (few-shot regression, image classification, reinforcement learning).
Achieves state-of-the-art performance on few-shot image classification benchmarks compared to specialized one-shot learning methods.
Shows that MAML accelerates reinforcement learning in the presence of task variability, outperforming direct pretraining as initialization.
Provides an algorithm that can be readily applied to regression and can accelerate reinforcement learning.

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.

Citation Graph

Loading graph...

References [40]

Sort:

Filter:

[1]Adam: A Method for Stochastic Optimization

D. P. Kingma, Jimmy Lei Ba - 2014

49 papers in library cite

Amazing paper! Very well explained and huge impact. I am amazed that they made something so simple even when it requires a lot of background mathematical knowledge

[2]Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

S. Ioffe, Christian Szegedy - 2015

18 papers in library cite

Very good paper! Similar feel as ResNets: simple idea, elegant. Not too mathy

[3]Explaining and Harnessing Adversarial Examples

Ian J. Goodfellow, J. Shlens, Christian Szegedy - 2015

4 papers in library cite

It feels that it is an extension of the previous paper, which is nice. It is less original but also easier to read!

[4]Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning

R. Williams - 1992

11 papers in library cite

It's alright for formalizing the concept, but it's a bit boring and doesn't add a lot from the middle on. Focuses too much in reviewing existing techniques and in stochastic units.

[5]TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems

M. Abadi, Akshat Agarwal, P. Barham, E. Brevdo, Ziru Chen, C. Citro, G. Corrado, A. Davis, Jeffrey Dean, M. Devin, Sanjay Ghemawat, I. Goodfellow, A. Harp, Geoffrey Irving, M. Isard, Y. Jia, R. Jozefowicz, Lukasz Kaiser, M. Kudlur, J. Levenberg, D. Mane, R. Monga, S. Moore, D. Murray, Christopher Olah, M. Schuster, J. Shlens, B. Steiner, Ilya Sutskever, K. Talwar, P. Tucker, Vincent Vanhoucke, V. Vasudevan, F. Viegas, Oriol Vinyals, P. Warden, M. Wattenberg, M. Wicke, Y. Yu, Xiaoqiang Zheng - 2015

11 papers in library cite

This should be the golden standard to what framework papers should be. It's large, but it's not boring at all. Explains the core concepts while not going too deep as to describe unimportant things; explains design decisions and shortcomings... overall amazing

[6]Overcoming Catastrophic Forgetting in Neural Networks

J. Kirkpatrick, Razvan Pascanu, N. C. Rabinowitz, J. Veness, G. Desjardins, A. A. Rusu, K. Milan, J. Quan, T. Ramalho, A. G. Barwinska, Demis Hassabis, C. Clopath, D. Kumaran, Raia Hadsell - 2017

5 papers in library cite

Very nice, intuitive, and impressive results.

[7]DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition

J. Donahue, Y. Jia, Oriol Vinyals, J. Hoffman, N. Zhang, E. Tzeng, Trevor Darrell - 2014

15 papers in library cite

Very nice paper. First I've seen (and based on the text, first ever) about feature extraction for images. It's very nice to see embeddings doing SotA

[8]Learning to Learn Using Gradient Descent

Sepp Hochreiter, A. Steven Younger, Peter R. Conwell - 2001

4 papers in library cite

I don't see how this is meta-learning.

[9]Exact Solutions to the Nonlinear Dynamics of Learning in Deep Linear Neural Networks

Surya Ganguli - 2014

9 papers in library cite

TBH it's been almost 2 months since I read this paper (shame on me for forgetting to add it)... Anyway, as I recall it I liked it, but TBH it's a bit underwhelming because it solved only for linear networks

[10]Learning a Synaptic Learning Rule

Yoshua Bengio, Samy Bengio, Jocelyn Cloutier - 1990

1 paper in library cites

WTF is this Bengio? An incomplete draft? Terrible.

[11]Trust Region Policy Optimization

John Schulman, Sergey Levine, P. Abbeel, Michael I. Jordan, P. Moritz - 2015

4 papers in library cite

[12]Matching Networks for One Shot Learning

Oriol Vinyals, C. Blundell, T. Lillicrap, Daan Wierstra - 2016

2 papers in library cite

[13]Siamese Neural Networks for One-Shot Image Recognition

G. Koch - 2015

1 paper in library cites

Interesting siamese network

[14]Optimization as a Model for Few-Shot Learning

S. Ravi, Hugo Larochelle - 2017

2 papers in library cite

[15]Meta-Learning With Memory-Augmented Neural Networks

Adam Santoro, S. Bartunov, M. Botvinick, Daan Wierstra, T. Lillicrap - 2016

2 papers in library cite

[16]Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks

T. Salimans, D. A. Kingma, D. P. Diederik - 2016

4 papers in library cite

Cited by a lot of people

[17]Learning to Learn

Sebastian Thrun, L. Pratt - 1998

3 papers in library cite

Book

[18]HyperNetworks

D. Ha, Andrew Dai, Quoc V. Le - 2016

3 papers in library cite

Using a network to generate weights for other network

[19]R12: Fast Reinforcement Learning via Slow Reinforcement Learning

Y. Duan, John Schulman, X. Chen, P. L. Bartlett, Ilya Sutskever, P. Abbeel - 2016

2 papers in library cite

[20]Learning to Optimize

K. Li, Jitendra Malik - 2017

1 paper in library cites

[21]Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning

E. Parisotto, Jimmy Lei Ba, Ruslan Salakhutdinov - 2015

2 papers in library cite

Multitask learning

[22]One-Shot Generalization in Deep Generative Models

D. J. Rezende, S. Mohamed, Ivo Danihelka, K. Gregor, Daan Wierstra - 2016

2 papers in library cite

Generative, better than DRAW

[23]Learning to Control Fast-Weight Memories: An Alternative to Dynamic Recurrent Networks

Jürgen Schmidhuber - 1992

4 papers in library cite

[24]Evolutionary Principles in Self-Referential Learning, or on Learning How to Learn: The Meta-Meta-... Hook

Jürgen Schmidhuber - 1987

3 papers in library cite

[25]Learning to Learn by Gradient Descent by Gradient Descent

M. Andrychowicz, M. Denil, S. Gomez, M. W. Hoffman, D. Pfau, T. Schaul, N. D. Freitas - 2016

3 papers in library cite

[26]Mujoco: A Physics Engine for Model-Based Control

E. Todorov, T. Erez, Y. Tassa - 2012

3 papers in library cite

[27]Benchmarking Deep Reinforcement Learning for Continuous Control

Y. Duan, X. Chen, R. Houthooft, R. Rein, John Schulman, P. Abbeel - 2016

2 papers in library cite

[28]Attentive Recurrent Comparators

Pranav Shyam, S. Gupta, A. Dukkipati - 2017

1 paper in library cites

[29]Data-Dependent Initializations of Convolutional Neural Networks

P. Krahenbuhl, C. Doersch, J. Donahue, Trevor Darrell - 2016

1 paper in library cites

[30]Fast Learning for Problem Classes Using Knowledge Based Network Initialization

M. Husken, C. Goerick - 2000

1 paper in library cites

[31]Gradient-Based Hyperparameter Optimization Through Reversible Learning

D. Maclaurin, David Duvenaud, R. A. Adams - 2015

1 paper in library cites

[32]Learning to Reinforcement Learn

J. X. Wang, Z. K. Nelson, D. Tirumala, H. Soyer, J. Z. Leibo, Rémi Munos, C. Blundell, D. Kumaran, M. Botvinick - 2016

1 paper in library cites

[33]Learning to Remember Rare Events

Lukasz Kaiser, O. Nachum, A. Roy, Samy Bengio - 2017

1 paper in library cites

[34]Meta Networks

T. Munkhdalai, H. Yu - 2017

1 paper in library cites

[35]Meta-Neural Networks That Learn by Learning

D. K. Naik, R. J. Mammone - 1992

1 paper in library cites

[36]On the Optimization of a Synaptic Learning Rule

Samy Bengio, Samy Bengio, Y. Yoshua, C. Jocelyn, J. Gecsei - 1992

1 paper in library cites

[37]One Shot Learning of Simple Visual Concepts

B. M. Lake, Ruslan Salakhutdinov, J. Gross, Joshua B. Tenenbaum - 2011

1 paper in library cites

[38]Online Representation Learning in Recurrent Neural Language Models

M. Rei - 2015

1 paper in library cites

[39]Prototypical Networks for Few-Shot Learning

J. Snell, K. Swersky, Richard S. Zemel - 2017

1 paper in library cites

[40]Towards a neural Statistician

Harri Edwards, A. Storkey - 2017

1 paper in library cites

Cited by

4

papers in your library

Cites

22

papers in your library

Read

on November 5, 2025

Very nice and very well explained. At first it seems that it will be tough to follow, but once they explain it it gets very intuitive.

Tags

Paper Aliases

No aliases