Few-Shot Learning with Graph Neural Networks

Abstract

We propose to study the problem of few-shot learning with the prism of inference on a partially observed graphical model, constructed from a collection of input images whose label can be either observed or not. By assimilating generic message-passing inference algorithms with their neural-network counterparts, we define a graph neural network architecture that generalizes several of the recently proposed few-shot learning models. Besides providing improved numerical performance, our framework is easily extended to variants of few-shot learning, such as semi-supervised or active learning, demonstrating the ability of graph-based models to operate well on 'relational' tasks.

References in corpus (5)

• Semi-Supervised Classification with Graph Convolutional Networks
• Deep Convolutional Networks on Graph-Structured Data
• Learning to Remember Rare Events
• A Compositional Object-Based Approach to Learning Physical Dynamics
• Generative Adversarial Residual Pairwise Networks for One Shot Learning

Cited by in corpus (169)

• Graph Neural Networks: A Review of Methods and Applications
• Graph Neural Network for Traffic Forecasting: A Survey
• Neural Relational Inference for Interacting Systems
• Revisiting Graph Neural Networks: All We Have is Low-Pass Filters
• SimpleShot: Revisiting Nearest-Neighbor Classification for Few-Shot Learning
• Graph signal processing for machine learning: A review and new perspectives
• A Closer Look at Few-shot Classification
• LanczosNet: Multi-Scale Deep Graph Convolutional Networks
• Deep Learning on Graphs: A Survey
• A Two-Stage Approach to Few-Shot Learning for Image Recognition
• Few-Shot Graph Learning for Molecular Property Prediction
• Meta-learning with differentiable closed-form solvers
• Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformation
• Few-Shot Segmentation Propagation with Guided Networks
• Machine Learning on Graphs: A Model and Comprehensive Taxonomy
• Finding Task-Relevant Features for Few-Shot Learning by Category Traversal
• Few-Shot Learning via Embedding Adaptation with Set-to-Set Functions
• Few-shot Relation Extraction via Bayesian Meta-learning on Relation Graphs
• CrossTransformers: spatially-aware few-shot transfer
• Small Sample Learning in Big Data Era
• Few-shot Text Classification with Distributional Signatures
• Graph Partition Neural Networks for Semi-Supervised Classification
• Revisiting Local Descriptor based Image-to-Class Measure for Few-shot Learning
• Leveraging Siamese Networks for One-Shot Intrusion Detection Model
• DeepEMD: Differentiable Earth Mover's Distance for Few-Shot Learning
• Graph Representation Learning for Road Type Classification
• Transductive Episodic-Wise Adaptive Metric for Few-Shot Learning
• Learning from Few Samples: A Survey
• Information Symmetry Matters: A Modal-Alternating Propagation Network for Few-Shot Learning
• Graphite: Iterative Generative Modeling of Graphs
• Prior-Knowledge and Attention-based Meta-Learning for Few-Shot Learning
• Negative Margin Matters: Understanding Margin in Few-shot Classification
• AGL: a Scalable System for Industrial-purpose Graph Machine Learning
• Leveraging the Feature Distribution in Transfer-based Few-Shot Learning
• Incremental Few-Shot Learning with Attention Attractor Networks
• Edge-labeling Graph Neural Network for Few-shot Learning
• Graph Neural Based End-to-end Data Association Framework for Online Multiple-Object Tracking
• Self-Supervised Learning For Few-Shot Image Classification
• Graph Neural Networks with Generated Parameters for Relation Extraction
• Part-aware Prototype Network for Few-shot Semantic Segmentation
• A Neural Topic-Attention Model for Medical Term Abbreviation Disambiguation
• DPGN: Distribution Propagation Graph Network for Few-shot Learning
• SLAPS: Self-Supervision Improves Structure Learning for Graph Neural Networks
• ROD: Reception-aware Online Distillation for Sparse Graphs
• Large Margin Few-Shot Learning
• Probabilistic Model-Agnostic Meta-Learning
• Weakly-supervised Compositional FeatureAggregation for Few-shot Recognition
• Image Deformation Meta-Networks for One-Shot Learning
• Adversarial Feature Hallucination Networks for Few-Shot Learning
• Automated Relational Meta-learning
• A Baseline for Few-Shot Image Classification
• Graph Element Networks: adaptive, structured computation and memory
• Multi-Level Matching and Aggregation Network for Few-Shot Relation Classification
• BERT Learns (and Teaches) Chemistry
• Cross-Domain Few-Shot Learning with Meta Fine-Tuning
• Modality to Modality Translation: An Adversarial Representation Learning and Graph Fusion Network for Multimodal Fusion
• Learning to Learn Variational Semantic Memory
• PeerNets: Exploiting Peer Wisdom Against Adversarial Attacks
• RelationNet2: Deep Comparison Columns for Few-Shot Learning
• Adaptive Deep Kernel Learning
• GIKT: A Graph-based Interaction Model for Knowledge Tracing
• Few-shot Action Recognition with Permutation-invariant Attention
• Evaluating Deep Graph Neural Networks
• Induction Networks for Few-Shot Text Classification
• Relational Generalized Few-Shot Learning
• A Meta-Learning Approach for Graph Representation Learning in Multi-Task Settings
• A Broader Study of Cross-Domain Few-Shot Learning
• Few-Shot Video Classification via Temporal Alignment
• Prototype Rectification for Few-Shot Learning
• Multi-scale Adaptive Task Attention Network for Few-Shot Learning
• Self-Supervision with Superpixels: Training Few-shot Medical Image Segmentation without Annotation
• Generalization Bounds For Meta-Learning: An Information-Theoretic Analysis
• Few-shot Classification via Adaptive Attention
• Reconciling meta-learning and continual learning with online mixtures of tasks
• SCG-Net: Self-Constructing Graph Neural Networks for Semantic Segmentation
• Deep Metric Transfer for Label Propagation with Limited Annotated Data
• Learning to Learn with Variational Information Bottleneck for Domain Generalization
• Learning event representations for temporal segmentation of image sequences by dynamic graph embedding
• Meta-Learning with Fewer Tasks through Task Interpolation
• FGN: Fully Guided Network for Few-Shot Instance Segmentation
• L2AE-D: Learning to Aggregate Embeddings for Few-shot Learning with Meta-level Dropout
• Region Comparison Network for Interpretable Few-shot Image Classification
• Context-Aware Visual Compatibility Prediction
• Learning to Affiliate: Mutual Centralized Learning for Few-shot Classification
• Learning Color Compatibility in Fashion Outfits
• Graph-based Interpolation of Feature Vectors for Accurate Few-Shot Classification
• ProtoTransformer: A Meta-Learning Approach to Providing Student Feedback
• An Attention-Based System for Damage Assessment Using Satellite Imagery
• Online Structured Meta-learning
• Learning Context Graph for Person Search
• Learning from the Past: Continual Meta-Learning via Bayesian Graph Modeling
• MM-FSOD: Meta and metric integrated few-shot object detection
• Sequential Graph Convolutional Network for Active Learning
• Data-Efficient GAN Training Beyond (Just) Augmentations: A Lottery Ticket Perspective
• Hierarchical Message-Passing Graph Neural Networks
• AgileNet: Lightweight Dictionary-based Few-shot Learning
• Attentive Graph Neural Networks for Few-Shot Learning
• Generalized Zero and Few-Shot Transfer for Facial Forgery Detection
• Weakly-supervised Object Localization for Few-shot Learning and Fine-grained Few-shot Learning
• Learning Global and Local Consistent Representations for Unsupervised Image Retrieval via Deep Graph Diffusion Networks
• Graph Representation learning for Audio & Music genre Classification
• One to Many: Adaptive Instrument Segmentation via Meta Learning and Dynamic Online Adaptation in Robotic Surgical Video
• Rethinking Class Relations: Absolute-relative Supervised and Unsupervised Few-shot Learning
• Tailoring: encoding inductive biases by optimizing unsupervised objectives at prediction time
• Compositional Few-Shot Recognition with Primitive Discovery and Enhancing
• FOAL: Fast Online Adaptive Learning for Cardiac Motion Estimation
• Learning Task-General Representations with Generative Neuro-Symbolic Modeling
• Boosting Supervision with Self-Supervision for Few-shot Learning
• Entity Concept-enhanced Few-shot Relation Extraction
• Continual Local Replacement for Few-shot Learning
• Revisiting Metric Learning for Few-Shot Image Classification
• Large-Scale Historical Watermark Recognition: dataset and a new consistency-based approach
• Dual Graph Convolutional Networks with Transformer and Curriculum Learning for Image Captioning
• Object-aware Long-short-range Spatial Alignment for Few-Shot Fine-Grained Image Classification
• Hierarchical Representation based Query-Specific Prototypical Network for Few-Shot Image Classification
• Contextualizing Enhances Gradient Based Meta Learning
• Revisiting Mid-Level Patterns for Cross-Domain Few-Shot Recognition
• Learning to See Through Obstructions with Layered Decomposition
• RICE: Refining Instance Masks in Cluttered Environments with Graph Neural Networks
• Improving Few-shot Learning by Spatially-aware Matching and CrossTransformer
• Are Fewer Labels Possible for Few-shot Learning?
• Graph Neural Networks for Image Understanding Based on Multiple Cues: Group Emotion Recognition and Event Recognition as Use Cases
• Learning Compositional Representation for Few-shot Visual Question Answering
• Meta Learning with Differentiable Closed-form Solver for Fast Video Object Segmentation
• Multi-step Estimation for Gradient-based Meta-learning
• Learning Predicates as Functions to Enable Few-shot Scene Graph Prediction
• Model-Agnostic Graph Regularization for Few-Shot Learning
• Hierarchical Meta Learning
• High-order structure preserving graph neural network for few-shot learning
• One-Click Annotation with Guided Hierarchical Object Detection
• Multi-Label Few-Shot Learning for Aspect Category Detection
• Deep Learning of Unified Region, Edge, and Contour Models for Automated Image Segmentation
• Few Shot Activity Recognition Using Variational Inference
• Data-Efficient Graph Embedding Learning for PCB Component Detection
• Bidirectional RNN-based Few Shot Learning for 3D Medical Image Segmentation
• AMP0: Species-Specific Prediction of Anti-microbial Peptides using Zero and Few Shot Learning
• Adaptive Prototypical Networks with Label Words and Joint Representation Learning for Few-Shot Relation Classification
• Dynamic Graph Modules for Modeling Object-Object Interactions in Activity Recognition
• Learning Implicit Temporal Alignment for Few-shot Video Classification
• Confusable Learning for Large-class Few-Shot Classification
• Graph convolutional networks for learning with few clean and many noisy labels
• Global Context Networks
• Personalized Hashtag Recommendation for Micro-videos
• Will Multi-modal Data Improves Few-shot Learning?
• Adaptive Knowledge-Enhanced Bayesian Meta-Learning for Few-shot Event Detection
• Few-shot learning via tensor hallucination
• Memory-Augmented Relation Network for Few-Shot Learning
• Tackling Early Sparse Gradients in Softmax Activation Using Leaky Squared Euclidean Distance
• Representation based and Attention augmented Meta learning
• Extended Few-Shot Learning: Exploiting Existing Resources for Novel Tasks
• Uncertainty-Aware Few-Shot Image Classification
• A Generative Approach Towards Improved Robotic Detection of Marine Litter
• Edge-Labeling based Directed Gated Graph Network for Few-shot Learning
• Meta-Active Learning for Node Response Prediction in Graphs
• Adaptation-Agnostic Meta-Training
• Learning Nonparametric Human Mesh Reconstruction from a Single Image without Ground Truth Meshes
• Shoestring: Graph-Based Semi-Supervised Learning with Severely Limited Labeled Data
• OR-Net: Pointwise Relational Inference for Data Completion under Partial Observation
• Domain Agnostic Few-Shot Learning For Document Intelligence
• Improving Text-Independent Speaker Verification with Auxiliary Speakers Using Graph
• Class Interference Regularization
• One Representative-Shot Learning Using a Population-Driven Template with Application to Brain Connectivity Classification and Evolution Prediction
• Decoupling Features and Coordinates for Few-shot RGB Relocalization
• Optimization of Image Embeddings for Few Shot Learning
• Ensemble Making Few-Shot Learning Stronger
• SAHDL: Sparse Attention Hypergraph Regularized Dictionary Learning
• CLTA: Contents and Length-based Temporal Attention for Few-shot Action Recognition
• External-Memory Networks for Low-Shot Learning of Targets in Forward-Looking-Sonar Imagery
• Self-Denoising Neural Networks for Few Shot Learning