Skip to content

Latest commit

 

History

History
279 lines (219 loc) · 42.2 KB

README.md

File metadata and controls

279 lines (219 loc) · 42.2 KB
Raw
👆 右上角点击 告诉我，你希望这个项目继续加速更新迭代 ❤️ & ☕️

Meta-Learning-Papers-with-Code

Awesome PRs Welcome Total Visitors Repo Clones Stars

This repository contains a reading list of papers with code on Meta-Learning and Meta-Reinforcement-Learning, These papers are mainly categorized according to the type of model. In addition, I will separately list papers from important conferences starting from 2023, e.g., NIPS, ICML, ICLR, CVPR etc. This repository is still being continuously improved. If you have found any relevant papers that need to be included in this repository, please feel free to submit a pull request (PR) or open an issue.

Each paper may be applicable to one or more types of meta-learning frameworks, including optimization-based and metric-based, and may be applicable to multiple data sources, including image, text, audio, video, and multi-modality. These are marked in the type column. In addition, for different tasks and different problems, we have marked the SOTA algorithm separately. This is submitted with reference to the leadboard at the time of submission, and will be continuously modified. We provide a basic introduction to each paper to help you understand the work and core ideas of this article more quickly.

Label

🎭 Different Frameworks

  • Meta-Learning Meta-Learning.
  • Meta-Reinforcement-Learning Meta-Reinforcement-Learning.

🎨 Different Types

  • optimization-based Optimization-based meta-learning approaches acquire a collection of optimal initial parameters, facilitating rapid convergence of a model when adapting to novel tasks.
  • metric-based Metric-based meta-learning approaches acquire embedding functions that transform instances from various tasks, allowing them to be readily categorized using non-parametric methods.

✨ Different Data Sources

  • Image Meta-Learning for CV (Images)
  • Video Meta-Learning for CV (Videos)
  • Text Meta-Learning for NLP
  • Audio Meta-Learning for Audio
  • Multi Meta-Learning for Multi-modal

It is worth noting that the experiments of some frameworks consist of multiple data sources. Our annotations are based on the paper description.

🎁 Notice

  • ❗CODE The paper does not provide code, I will write it myself and supplement it later.
  • ❗Oral The Oral paper.
  • ❗Spotlight The Oral paper.

🚩 I have marked some recommended papers with 🌟/🎈 (SOTA methods/Just my personal preference 😉).

🚩 I will maintain three hours of paper reading, code repository maintenance and entry supplement every day 😉).

🚩 All papers are provided in the corresponding folders 😉.

Topics

Survey.

Date Method Type Conference Paper Title and Paper Interpretation Code
2018 RL L2L Meta-Reinforcement-Learning arXiv 2018 A review of meta-reinforcement learning for deep neural networks architecture search None
2019 Book of Meta-Learning Meta-Learning Book Meta-Learning (Automated Machine Learning) None
2019 Learn dynamics Meta-Learning arXiv 2019 Meta-learners' learning dynamics are unlike learners' None
2020 NLP🌟 Meta-Learning arXiv 2020 Meta-learning for few-shot natural language processing: A survey None
2020 CV-classifier Meta-Learning IEEE Access A literature survey and empirical study of meta-learning for classifier selection None
2020 RL DL L2L Meta-Learning Meta-Reinforcement-Learning arXiv 2020 A comprehensive overview and survey of recent advances in meta-learning None
2021 Learn 2 Learn Meta-Learning arXiv 2021 Meta-Learning: A Survey None
2021 Learn 2 Learn 🎈 Meta-Learning TPAMI Meta-Learning in Neural Networks: A Survey None
2021 Learn 2 Learn Meta-Learning Artif Intell Rev A survey of deep meta-learning None
2021 Learn 2 Learn Meta-Learning Current Opinion in Behavioral Sciences Meta-learning in natural and artificial intelligence None
2022 Multi-Modal🌟 Meta-Learning KBS Multimodality in meta-learning: A comprehensive survey None
2022 Image Segmentation🌟 Meta-Learning PR Meta-seg: A survey of meta-learning for image segmentation None
2022 Cyberspace Security Meta-Learning Digit. Commun. Netw. Application of meta-learning in cyberspace security: A survey None
2023 RL L2L🌟 Meta-Reinforcement-Learning arXiv 2023 A survey of meta-reinforcement learning None

Optimization

Date Method Type Conference Paper Title and Paper Interpretation Code
2016 Reversible Meta-Learning optimization-based Image ICML 2016 Gradient-based Hyperparameter Optimization through Reversible Learning CODE
2017 MRL-GPS Meta-Reinforcement-Learning optimization-based ICLR 2017 Learning to Optimize ❗CODE
2019 L2G Meta-Learning metric-based Image arXiv 2019 Learning to Generalize to Unseen Tasks with Bilevel Optimization ❗CODE
2019 LOIS Meta-Learning optimization-based arXiv 2019 Learning to Optimize in Swarms CODE
2019 iMAML🌟 Meta-Learning optimization-based NIPS 2019 Meta-Learning with Implicit Gradients CODE
2019 Xfer🌟 Meta-Learning optimization-based ICLR 2019 Transferring Knowledge across Learning Processes CODE
2019 MetaInit Meta-Learning optimization-based ICLR 2019 MetaInit: Initializing learning by learning to initialize ❗CODE
2019 Runge-Kutta-MAML Meta-Learning optimization-based arXiv 2019 MetaInit: Initializing learning by learning to initialize ❗CODE
2020 WarpGrad Meta-Learning optimization-based ICLR 2020 Model-Agnostic Meta-Learning using Runge-Kutta Methods CODE
2022 Sharp-MAML🎈 Meta-Learning optimization-based Image ICML 2022 Sharp-MAML: Sharpness-Aware Model-Agnostic Meta Learning CODE
2022 BMG🌟 Meta-Learning optimization-based Image ICLR 2022 Bootstrapped Meta-Learning ❗CODE

Theory

Date Method Type Conference Paper Title and Paper Interpretation Code
2018 MLAP Meta-Learning optimization-based ICML 2018 Meta-learning by adjusting priors based on extended PAC-Bayes theory CODE
2018 learning algorithm approximation Meta-Learning optimization-based ICLR 2018 Meta-Learning and Universality: Deep Representations and Gradient Descent can Approximate any Learning Algorithm ❗CODE
2018 ConsiderMRL Meta-Reinforcement-Learning optimization-based ICLR 2018 Some Considerations on Learning to Explore via Meta-Reinforcement Learning CODE
2022 UMAML Meta-Learning optimization-based ICLR 2022 Unraveling Model-Agnostic Meta-Learning via The Adaptation Learning Rate ❗CODE
2022 TRGB Meta-Learning optimization-based ICLR 2022 Task Relatedness-Based Generalization Bounds for Meta Learning ❗CODE
2021 PAC-Bayes Meta-Learning optimization-based NeurIPS 2021 How Tight Can PAC-Bayes be in the Small Data Regime? ❗CODE
2021 meta_tr_val_split Meta-Learning metric-based ICML 2021 A Representation Learning Perspective on the Importance of Train-Validation Splitting in Meta-Learning CODE
2021 stocBiO Meta-Learning optimization-basedImage ICML 2021 Bilevel Optimization: Convergence Analysis and Enhanced Design CODE
2022 First active ML Meta-Learning optimization-based AISTATS 2022 Near-Optimal Task Selection with Mutual Information for Meta-Learning ❗CODE
2022 LTR Meta-Learning optimization-based AISTATS 2022 Learning Tensor Representations for Meta-Learning ❗CODE
2022 BayesianMAML or MAML? Meta-Learning optimization-based AISTATS 2022 Is Bayesian Model-Agnostic Meta Learning Better than Model-Agnostic Meta Learning, Provably? ❗CODE
2023 ​​What learning algorithm is in-context learning? Meta-Learning ICLR 2023 ​​What learning algorithm is in-context learning? Investigations with linear models❗Notable-top-5Percent CODE

Domain generalization

Date Method Type Conference Paper Title and Paper Interpretation Code
2018 L2G Meta-Learning Meta-Reinforcement-Learning optimization-based AAAI 2018 Learning to Generalize: Meta-Learning for Domain Generalization CODE
2019 MASF Meta-Learning optimization-based metric-based NIPS 2019 Domain Generalization via Model-Agnostic Learning of Semantic Features CODE
2020 MLCA Meta-Reinforcement-Learning ICLR 2020 Meta-learning curiosity algorithms CODE

Lifelong learning

Date Method Type Conference Paper Title and Paper Interpretation Code
2018 IL2L🌟 Meta-Learning optimization-based arXiv 2018 Incremental Learning-to-Learn with Statistical Guarantees ❗CODE
2019 VividNet Meta-Learning Graph arXiv 2019 A Neural-Symbolic Architecture for Inverse Graphics Improved by Lifelong Meta-Learning CODE
2019 HSML Meta-Learning optimization-based Image Text ICML 2019 Hierarchically Structured Meta-learning CODE
2019 Online-ML🌟 Meta-Learning optimization-based ICML 2019 Online Meta-Learning ❗CODE
2019 MRCL Meta-Learning optimization-based NIPS 2019 Meta-Learning Representations for Continual Learning CODE
2019 Bayes-MAML Meta-Learning optimization-based NIPS 2019 Reconciling meta-learning and continual learning with online mixtures of tasks ❗CODE
2019 ONL-ONL Meta-Learning optimization-based NIPS 2019 Online-Within-Online Meta-Learning CODE
2021 LWTL🌟 Meta-Learning optimization-based NIPS 2021 Learning where to learn: Gradient sparsity in meta and continual learning CODE
2021 MARK🌟 Meta-Learning optimization-basedVideo NIPS 2021 Optimizing Reusable Knowledge for Continual Learning via Metalearning CODE

Model compression

Summary of conference papers

CVPR23

Date Method Type Conference Paper Title and Paper Interpretation Code
2023 PPL🌟 Meta-Learning optimization-based CVPR 2023 A Meta-Learning Approach to Predicting Performance and Data Requirements ❗CODE
2023 Meta-Explore Meta-Learning optimization-based CVPR 2023 Meta-Explore: Exploratory Hierarchical Vision-and-Language Navigation Using Scene Object Spectrum Grounding ❗CODE
2023 Meta-Tuning🌟 Meta-Learning optimization-based CVPR 2023 Meta-Tuning Loss Functions and Data Augmentation for Few-Shot Object Detection ❗CODE
2023 Meta-Causal-learning🌟 Meta-Learning optimization-based CVPR 2023 Meta-Tuning Loss Functions and Data Augmentation for Few-Shot Object Detection ❗CODE
2023 Model-Scale Agnostic🌟 Meta-Learning optimization-based CVPR 2023 Architecture, Dataset and Model-Scale Agnostic Data-free Meta-Learning. CODE

ICML23

Date Method Type Conference Paper Title and Paper Interpretation Code
2023 MLPS🌟 Meta-Reinforcement-Learning optimization-based ICML 2023 Meta-Learning Parameterized Skills CODE
2023 Meta-Meta-Learning Meta-Learning optimization-based ICML 2023 Learning to acquire novel cognitive tasks with evolution, plasticity and meta-meta-learning CODE
2023 BiDf-MKD🌟 Meta-Learning optimization-based ICML 2023 Learning to Learn from APIs: Black-Box Data-Free Meta-Learning CODE
2023 Meta-SAGE Meta-Reinforcement-Learning ICML 2023 Meta-SAGE: Scale Meta-Learning Scheduled Adaptation with Guided Exploration for Mitigating Scale Shift on Combinatorial Optimization CODE
2023 RepVerb Meta-Learningoptimization-basedTextImage ICML 2023 Effective Structured Prompting by Meta-Learning and Representative Verbalizer ❗CODE
2023 Memory-Based Meta-Learning Meta-Learningoptimization-basedVideo ICML 2023 Memory-Based Meta-Learning on Non-Stationary Distributions CODE

NIPS23

Date Method Type Conference Paper Title and Paper Interpretation Code
2023 Conformal-Meta🌟 Meta-Learning optimization-based NIPS 2023 Conformal Meta-learners for Predictive Inference of Individual Treatment Effects ❗Oral CODE
2023 MGDD Meta-Learning NIPS 2023 Online Constrained Meta-Learning: Provable Guarantees for Generalization ❗Spotlight ❗CODE
2023 PINNs Meta-Learning optimization-based NIPS 2023 MGDD: A Meta Generator for Fast Dataset Distillation ❗Spotlight ❗CODE
2023 OCML Meta-Learning optimization-based NIPS 2023 Hypernetwork-based Meta-Learning for Low-Rank Physics-Informed Neural Networks ❗Spotlight ❗CODE
2023 Online Control Meta-Learning optimization-based NIPS 2023 Online Control for Meta-optimization ❗Spotlight ❗CODE
2023 SCARF Meta-Learning optimization-based NIPS 2023 Prefix-Tree Decoding for Predicting Mass Spectra from Molecules ❗Spotlight CODE
2023 HNPs Meta-Learning optimization-based NIPS 2023 Learning from Active Human Involvement through Proxy Value Propagation ❗Spotlight ❗CODE
2023 Zero-shot causal learning Meta-Learning optimization-based NIPS 2023 Episodic Multi-Task Learning with Heterogeneous Neural Processes ❗Spotlight CODE
2023 Zero-shot causal learning Meta-Learning optimization-based NIPS 2023 Zero-shot causal learning ❗Spotlight ❗CODE
2023 Structure-free Graph Condensation Meta-Learning optimization-based NIPS 2023 Structure-free Graph Condensation: From Large-scale Graphs to Condensed Graph-free Data ❗Spotlight ❗CODE
2023 Pick-up-to-Learn Meta-Learning optimization-based NIPS 2023 The Pick-to-Learn Algorithm: Empowering Compression for Tight Generalization Bounds and Improved Post-training Performance ❗Spotlight ❗CODE
2023 SimFBO Meta-Learning optimization-based NIPS 2023 SimFBO: Towards Simple, Flexible and Communication-efficient Federated Bilevel Learning ❗Spotlight ❗CODE
2023 EmbodiedGPT Meta-Learning optimization-based NIPS 2023 EmbodiedGPT: Vision-Language Pre-Training via Embodied Chain of Thought ❗Spotlight CODE

ICLR23

Date Method Type Conference Paper Title and Paper Interpretation Code
2023 Transfer NAS Meta-Learning optimization-based ICLR 2023 Transfer NAS with Meta-learned Bayesian Surrogates ❗NotableTop5Percent ❗CODE
2023 Betty🌟 Meta-Learning optimization-based ICLR 2023 Betty: An Automatic Differentiation Library for Multilevel Optimization ❗NotableTop5Percent CODE
2023 ​​What learning algorithm is in-context learning? Meta-Learning ICLR 2023 ​​What learning algorithm is in-context learning? Investigations with linear models ❗NotableTop5Percent CODE
2023 Learnable Behavior Control🌟 Meta-LearningMeta-Reinforcement-Learning ICLR 2023 Learnable Behavior Control: Breaking Atari Human World Records via Sample-Efficient Behavior Selection!❗NotableTop5Percent ❗CODE
2023 Metadata Archaeology Meta-Learning ICLR 2023 Metadata Archaeology: Unearthing Data Subsets by Leveraging Training Dynamics ❗NotableTop25Percent ❗CODE
2023 CMDP-within-online Meta-LearningMeta-Reinforcement-Learning ICLR 2023 A CMDP-within-online framework for Meta-Safe Reinforcement Learning ❗NotableTop25Percent ❗CODE
2023 MARS Meta-Learningoptimization-based ICLR 2023 MARS: Meta-learning as Score Matching in the Function Space ❗NotableTop25Percent ❗CODE

Libraries

Link
Awesome-META+
Higher by Facebook research
TorchMeta
Learn2learn

Blogs

Link
Multiple Meta-learning Papers
Berkeley Artificial Intelligence Research blog
Meta-Learning: Learning to Learn Fast
Meta-Reinforcement Learning
How to train your MAML: A step by step approach
An Introduction to Meta-Learning
From zero to research — An introduction to Meta-learning
What’s New in Deep Learning Research: Understanding Meta-Learning
Meta Reinforcement Learning Blog by Lilian Weng

Lecture Videos

Link
Stanford CS330: Multi-Task and Meta-Learning, 2019 by Chelsea Finn
Meta Learning lecture by Soheil Feizi
Chelsea Finn: Building Unsupervised Versatile Agents with Meta-Learning
Sam Ritter: Meta-Learning to Make Smart Inferences from Small Data
Model Agnostic Meta Learning by Siavash Khodadadeh
Meta Learning by Siraj Raval
Meta Learning by Hugo Larochelle
Meta Learning and One-Shot Learning

Datasets

Link
Omniglot
mini-ImageNet
ILSVRC
FGVC aircraft
Caltech-UCSD Birds-200-2011
Check several other datasets by Google here.

Workshops

Link
MetaLearn 2017
MetaLearn 2018
MetaLearn 2019
MetaLearn 2020

Researchers

Link
Chelsea Finn, UC Berkeley
Pieter Abbeel, UC Berkeley
Erin Grant, UC Berkeley
Raia Hadsell, DeepMind
Misha Denil, DeepMind
Adam Santoro, DeepMind
Sachin Ravi, Princeton University
David Abel, Brown University
Brenden Lake, Facebook AI Research