Meta-Learning: A Survey

Abstract

Meta-learning, or learning to learn, is the science of systematically observing how different machine learning approaches perform on a wide range of learning tasks, and then learning from this experience, or meta-data, to learn new tasks much faster than otherwise possible. Not only does this dramatically speed up and improve the design of machine learning pipelines or neural architectures, it also allows us to replace hand-engineered algorithms with novel approaches learned in a data-driven way. In this chapter, we provide an overview of the state of the art in this fascinating and continuously evolving field.

Cited by in corpus (84)

• Hyper-Parameter Optimization: A Review of Algorithms and Applications
• Fast Adaptive Task Offloading in Edge Computing based on Meta Reinforcement Learning
• Generalized Out-of-Distribution Detection: A Survey
• A Survey of Machine Learning for Computer Architecture and Systems
• Automated Machine Learning: State-of-The-Art and Open Challenges
• Predicting Water Temperature Dynamics of Unmonitored Lakes with Meta Transfer Learning
• Finetuned Language Models Are Zero-Shot Learners
• Meta-learning for Few-shot Natural Language Processing: A Survey
• Towards Green Automated Machine Learning: Status Quo and Future Directions
• Safe Reinforcement Learning via Curriculum Induction
• On the Efficiency of K-Means Clustering: Evaluation, Optimization, and Algorithm Selection
• Autonomous Driving with Deep Learning: A Survey of State-of-Art Technologies
• MetaMT,a MetaLearning Method Leveraging Multiple Domain Data for Low Resource Machine Translation
• Warm Up Cold-start Advertisements: Improving CTR Predictions via Learning to Learn ID Embeddings
• Weight-Sharing Neural Architecture Search: A Battle to Shrink the Optimization Gap
• Low-Cost Outdoor Air Quality Monitoring and Sensor Calibration: A Survey and Critical Analysis
• Meta-DETR: Image-Level Few-Shot Object Detection with Inter-Class Correlation Exploitation
• MetaIQA: Deep Meta-learning for No-Reference Image Quality Assessment
• Deep Learning to Estimate Permeability using Geophysical Data
• Meta-Learning GNN Initializations for Low-Resource Molecular Property Prediction
• Revisiting Meta-Learning as Supervised Learning
• How to Train Your MAML to Excel in Few-Shot Classification
• Fast-adapting and Privacy-preserving Federated Recommender System
• Physics-integrated machine learning: embedding a neural network in the Navier-Stokes equations. Part I
• Physics-integrated machine learning: embedding a neural network in the Navier-Stokes equations. Part II
• MAMO: Memory-Augmented Meta-Optimization for Cold-start Recommendation
• A Close Look at Deep Learning with Small Data
• Orchestrating the Development Lifecycle of Machine Learning-Based IoT Applications: A Taxonomy and Survey
• MetaCI: Meta-Learning for Causal Inference in a Heterogeneous Population
• Automating Outlier Detection via Meta-Learning
• An ADMM Based Framework for AutoML Pipeline Configuration
• Learning to Tune XGBoost with XGBoost
• Magnification Generalization for Histopathology Image Embedding
• Stateless Neural Meta-Learning using Second-Order Gradients
• CN-Celeb: multi-genre speaker recognition
• Artificial Intelligence in Drug Discovery: Applications and Techniques
• ProtoTransformer: A Meta-Learning Approach to Providing Student Feedback
• Data Augmentation for Object Detection via Differentiable Neural Rendering
• Amazon SageMaker Autopilot: a white box AutoML solution at scale
• Meta-Weighted Gaussian Process Experts for Personalized Forecasting of AD Cognitive Changes
• Lesion2Vec: Deep Metric Learning for Few-Shot Multiple Lesions Recognition in Wireless Capsule Endoscopy Video
• AutoML to Date and Beyond: Challenges and Opportunities
• Zero-shot Task Adaptation using Natural Language
• How can AI Automate End-to-End Data Science?
• Embedding Adaptation is Still Needed for Few-Shot Learning
• MetaDelta: A Meta-Learning System for Few-shot Image Classification
• Meta Automatic Curriculum Learning
• Learning to Discretize: Solving 1D Scalar Conservation Laws via Deep Reinforcement Learning
• Efficient AutoML Pipeline Search with Matrix and Tensor Factorization
• High-level Features for Resource Economy and Fast Learning in Skill Transfer
• TeachMyAgent: a Benchmark for Automatic Curriculum Learning in Deep RL
• Sign-MAML: Efficient Model-Agnostic Meta-Learning by SignSGD
• Distributed Evolution Strategies Using TPUs for Meta-Learning
• Domain-Invariant Speaker Vector Projection by Model-Agnostic Meta-Learning
• Automated Machine Learning Techniques for Data Streams
• Global Convergence and Generalization Bound of Gradient-Based Meta-Learning with Deep Neural Nets
• Meta-KD: A Meta Knowledge Distillation Framework for Language Model Compression across Domains
• Meta Fine-Tuning Neural Language Models for Multi-Domain Text Mining
• ALPaCA vs. GP-based Prior Learning: A Comparison between two Bayesian Meta-Learning Algorithms
• Fast Adaptation with Meta-Reinforcement Learning for Trust Modelling in Human-Robot Interaction
• How Low Can We Go: Trading Memory for Error in Low-Precision Training
• Solving Constrained CASH Problems with ADMM
• Federated Imitation Learning: A Novel Framework for Cloud Robotic Systems with Heterogeneous Sensor Data
• Prior-Enhanced Few-Shot Segmentation with Meta-Prototypes
• Hyperparameter Transfer Across Developer Adjustments
• Combining Domain-Specific Meta-Learners in the Parameter Space for Cross-Domain Few-Shot Classification
• A Unified Cognitive Learning Framework for Adapting to Dynamic Environment and Tasks
• Meta-Reinforcement Learning for Heuristic Planning
• Meta Cross-Modal Hashing on Long-Tailed Data
• Meta-Learning for One-Class Classification with Few Examples using Order-Equivariant Network
• Personalized Adaptive Meta Learning for Cold-start User Preference Prediction
• Meta-Reinforcement Learning for Trajectory Design in Wireless UAV Networks
• When Autonomous Systems Meet Accuracy and Transferability through AI: A Survey
• The Sample Complexity of Meta Sparse Regression
• Ensemble Making Few-Shot Learning Stronger
• Learning Abstract Task Representations
• Meta Learning for Support Recovery in High-dimensional Precision Matrix Estimation
• Crowdsourcing with Meta-Workers: A New Way to Save the Budget
• An Extensive Experimental Evaluation of Automated Machine Learning Methods for Recommending Classification Algorithms (Extended Version)
• A Semi-Personalized System for User Cold Start Recommendation on Music Streaming Apps
• Using Known Information to Accelerate HyperParameters Optimization Based on SMBO
• Transformative Machine Learning
• Towards Dynamic Feature Selection with Attention to Assist Banking Customers in Establishing a New Business
• Beyond Categorical Label Representations for Image Classification