paper-with-me

홈 › Papers

When MAML Can Adapt Fast and How to Assist When It Cannot

2019-10-30 · Sébastien M. R. Arnold, Shariq Iqbal, Fei Sha

Model-Agnostic Meta-Learning (MAML) and its variants have achieved success in meta-learning tasks on many datasets and settings. On the other hand, we have just started to understand and analyze how they are able to adapt fast to new tasks. For example, one popular hypothesis is that the algorithms learn good representations for transfer, as in multi-task learning. In this work, we contribute by providing a series of empirical and theoretical studies, and discover several interesting yet previously unknown properties of the algorithm. We find MAML adapts better with a deep architecture even if the tasks need only a shallow one (and thus, no representation learning is needed). While echoing previous findings by others that the bottom layers in deep architectures enable representation learning, we also find that upper layers enable fast adaptation by being meta-learned to perform adaptive gradient update when generalizing to new tasks. Motivated by these findings, we study several meta-optimization approaches and propose a new one for learning to optimize adaptively. Those approaches attain stronger performance in meta-learning both shallower and deeper architectures than MAML.

📄 PDF Abstract BibTeX arXiv:1910.13603

Code (1)

learnables/learn2learn 공식 구현 pytorch

Tasks

Meta-LearningMulti-Task LearningRepresentation Learning

Methods 이 논문이 사용한 방법론

MAML 설명 없음

Similar Papers 제목 키워드 기반

Performance-Weighed Policy Sampling for Meta-Reinforcement Learning

2020-12-10 · Ibrahim Ahmed, Marcos Quinones-Grueiro, Gautam Biswas

This paper discusses an Enhanced Model-Agnostic Meta-Learning (E-MAML) algorithm that generates fast convergence of the policy function from a small number of training examples when applied to new learning tasks. Built o…

Meta-LearningMeta Reinforcement Learningreinforcement-learningReinforcement Learning+1

Meta-Learning with Adaptive Hyperparameters

2020-10-31 · NeurIPS 2020 12 · Sungyong Baik, Myungsub Choi, Janghoon Choi, Heewon Kim 외

Despite its popularity, several recent works question the effectiveness of MAML when test tasks are different from training tasks, thus suggesting various task-conditioned methodology to improve the initialization. Inste…

Few-Shot LearningMeta-Learning

How Does the Task Landscape Affect MAML Performance?

2020-10-27 · Liam Collins, Aryan Mokhtari, Sanjay Shakkottai

Model-Agnostic Meta-Learning (MAML) has become increasingly popular for training models that can quickly adapt to new tasks via one or few stochastic gradient descent steps. However, the MAML objective is significantly m…

Few-Shot Image Classificationimage-classificationImage ClassificationMeta-Learning

On Fast Adversarial Robustness Adaptation in Model-Agnostic Meta-Learning

2021-02-20 · ICLR 2021 1 · Ren Wang, Kaidi Xu, Sijia Liu, Pin-Yu Chen 외

Model-agnostic meta-learning (MAML) has emerged as one of the most successful meta-learning techniques in few-shot learning. It enables us to learn a meta-initialization} of model parameters (that we call meta-model) to …

Adversarial AttackAdversarial RobustnessContrastive LearningData Augmentation+2

ES-MAML: Simple Hessian-Free Meta Learning

2019-09-25 · ICLR 2020 1 · Xingyou Song, Wenbo Gao, Yuxiang Yang, Krzysztof Choromanski 외

We introduce ES-MAML, a new framework for solving the model agnostic meta learning (MAML) problem based on Evolution Strategies (ES). Existing algorithms for MAML are based on policy gradients, and incur significant diff…

Meta-Learning