mlsanityData to 14 Oct 2026

On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features

arXiv 2111.12128 ↗First posted 23 Nov 2021cs.LGcited by 63

Not on the map. Why?

Abstract (arXiv)

While Graph Neural Networks (GNNs) have recently become the de facto standard for modeling relational data, they impose a strong assumption on the availability of the node or edge features of the graph. In many real-world applications, however, features are only partially available; for example, in social networks, age and gender are available only for a small subset of users. We present a general approach for handling missing features in graph machine learning applications that is based on minimization of the Dirichlet energy and leads to a diffusion-type differential equation on the graph. The discretization of this equation produces a simple, fast and scalable algorithm which we call Feature Propagation. We experimentally show that the proposed approach outperforms previous methods on seven common node-classification benchmarks and can withstand surprisingly high rates of missing features: on average we observe only around 4% relative accuracy drop when 99% of the features are missing. Moreover, it takes only 10 seconds to run on a graph with ~2.5M nodes and ~123M edges on a single GPU.

Key facts. Cited by 63 papers in the map (data through 14 Oct 2026). Builds on 35 of its 62 references that are in the map. First posted on arXiv 23 Nov 2021.

Builds on

35 of 62 references.

References · 35

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  3. 2017Graph Attention Networkscited by 8,923
  4. 2017Inductive Representation Learning on Large Graphscited by 6,320
  5. 2013Auto-Encoding Variational Bayescited by 4,053
  6. 2017Neural Message Passing for Quantum Chemistrycited by 3,487
  7. 2019Fast Graph Representation Learning with PyTorch Geometriccited by 2,434
  8. 2020Open Graph Benchmark: Datasets for Machine Learning on Graphscited by 2,163
  9. 2016node2vec: Scalable Feature Learning for Networkscited by 1,918
  10. 2014DeepWalk: Online Learning of Social Representationscited by 1,547
  11. 2019Simplifying Graph Convolutional Networkscited by 1,477
  12. 2018Relational inductive biases, deep learning, and graph networkscited by 1,360
23 more
  1. 2018Graph Convolutional Neural Networks for Web-Scale Recommender Systemscited by 1,138
  2. 2016Geometric deep learning: going beyond Euclidean datacited by 1,048
  3. 2018Pitfalls of Graph Neural Network Evaluationcited by 1,007
  4. 2020Learning to Simulate Complex Physics with Graph Networkscited by 880
  5. 2018Representation Learning on Graphs with Jumping Knowledge Networkscited by 877
  6. 2015Convolutional Networks on Graphs for Learning Molecular Fingerprintscited by 754
  7. 2020Benchmarking Graph Neural Networkscited by 680
  8. 2019MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixingcited by 443
  9. 2019Diffusion Improves Graph Learningcited by 414
  10. 2018GAIN: Missing Data Imputation using Generative Adversarial Netscited by 352
  11. 2021GRAND: Graph Neural Diffusioncited by 258
  12. 2018Modeling polypharmacy side effects with graph convolutional networkscited by 237
  13. 2020The Surprising Power of Graph Neural Networks with Random Node Initializationcited by 198
  14. 2020Random Features Strengthen Graph Neural Networkscited by 186
  15. 2019Continuous Graph Neural Networkscited by 114
  16. 2017Geometric Matrix Completion with Recurrent Multi-Graph Neural Networkscited by 88
  17. 2020Scalable Graph Neural Networks via Bidirectional Propagationcited by 87
  18. 2021On Positional and Structural Node Features for Graph Neural Networks on Non-attributed Graphscited by 49
  19. 2020Graph Convolutional Networks for Graphs Containing Missing Featurescited by 47
  20. 2020Learning on Attribute-Missing Graphscited by 42
  21. 2014Matrix Completion on Graphscited by 26
  22. 2020Incomplete Graph Representation and Learning via Partial Graph Neural Networkscited by 16
  23. 2021Propagation on Multi-relational Graphs for Node Regressioncited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 1; 2022: 6; 2023: 20; 2024: 9; 2025: 15; 2026: 12.

63 papers.

Sort:

By area:

  1. 2022Data Augmentation for Deep Graph Learning: A Surveycited by 133
  2. 2022Learning to Reconstruct Missing Data from Spatiotemporal Graphs with Sparse Observationscited by 64
  3. 2022A Survey on Fairness for Machine Learning on Graphscited by 20
  4. 2023Learnable Topological Features for Phylogenetic Inference via Graph Neural Networkscited by 14
  5. 2023Confidence-Based Feature Imputation for Graphs with Partially Known Featurescited by 14
  6. 2023TouchUp-G: Improving Feature Representation through Graph-Centric Finetuningcited by 12
  7. 2022Recent Advances in Reliable Deep Graph Learning: Inherent Noise, Distribution Shift, and Adversarial Attackcited by 10
  8. 2022EGG-GAE: scalable graph neural networks for tabular data imputationcited by 7
  9. 2023Tractable Probabilistic Graph Representation Learning with Graph-Induced Sum-Product Networkscited by 7
  10. 2023Does Black-box Attribute Inference Attacks on Graph Neural Networks Constitute Privacy Risk?cited by 7
  11. 2023GRAFENNE: Learning on Graphs with Heterogeneous and Dynamic Feature Setscited by 7
  12. 2023Supercharging Graph Transformers with Advective Diffusioncited by 7
38 more
  1. 2023Missing Value Imputation for Multi-attribute Sensor Data Streams via Message Propagation (Extended Version)cited by 7
  2. 2024Spatiotemporal Implicit Neural Representation as a Generalized Traffic Data Learnercited by 7
  3. 2023Uncertainty-Aware Robust Learning on Noisy Graphscited by 6
  4. 2023Efficient Anatomical Labeling of Pulmonary Tree Structures via Deep Point-Graph Representation-based Implicit Fieldscited by 6
  5. 2025Toward General and Robust LLM-enhanced Text-attributed Graph Learningcited by 6
  6. 2023Recovering Missing Node Features with Local Structure-based Embeddingscited by 5
  7. 2023AmGCL: Feature Imputation of Attribute Missing Graph via Self-supervised Contrastive Learningcited by 4
  8. 2025Network-wide Freeway Traffic Estimation Using Sparse Sensor Data: A Dirichlet Graph Auto-Encoder Approachcited by 4
  9. 2023Self-supervised Guided Hypergraph Feature Propagation for Semi-supervised Classification with Missing Node Featurescited by 3
  10. 2023Enhanced Graph Neural Networks with Ego-Centric Spectral Subgraph Embeddings Augmentationcited by 3
  11. 2025Incomplete Graph Learning: A Comprehensive Surveycited by 3
  12. 2026Spatiotemporal Imputation with Graph-Informed Flow Matchingcited by 3
  13. 2023TwERC: High Performance Ensembled Candidate Generation for Ads Recommendation at Twittercited by 2
  14. 2024Gradient Flow of Energy: A General and Efficient Approach for Entity Alignment Decodingcited by 2
  15. 2024Augmenting Textual Generation via Topology Aware Retrievalcited by 2
  16. 2025Graph Your Own Promptcited by 2
  17. 2026Rethinking GNNs and Missing Features: Challenges, Evaluation and a Robust Solutioncited by 2
  18. 2021Designing weighted and multiplex networks for deep learning user geolocation in Twittercited by 1
  19. 2023Towards Unsupervised Graph Completion Learning on Graphs with Features and Structure Missingcited by 1
  20. 2023Variational Graph Autoencoder for Heterogeneous Information Networks with Missing and Inaccurate Attributescited by 1
  21. 2024Full-Body Motion Reconstruction with Sparse Sensing from Graph Perspectivecited by 1
  22. 2025Learning Cell-Aware Hierarchical Multi-Modal Representations for Robust Molecular Modelingcited by 1
  23. 2022Remote Work Optimization with Robust Multi-channel Graph Neural Networkscited by 0
  24. 2023Using simulation to calibrate real data acquisition in veterinary medicinecited by 0
  25. 2023Fair Graph Machine Learning under Adversarial Missingness Processescited by 0
  26. 2023A Consistent Diffusion-Based Algorithm for Semi-Supervised Graph Learningcited by 0
  27. 2024DPGAN: A Dual-Path Generative Adversarial Network for Missing Data Imputation in Graphscited by 0
  28. 2024A Spectral Framework for Evaluating Geodesic Distances Between Graphscited by 0
  29. 2024On zero-shot learning in neural state estimation of power distribution systemscited by 0
  30. 2024Convolutional Signal Propagation: A Simple Scalable Algorithm for Hypergraphscited by 0
  31. 2024Modelling Networked Dynamical System by Temporal Graph Neural ODE with Irregularly Partial Observed Time-series Datacited by 0
  32. 2025Topology-Driven Attribute Recovery for Attribute Missing Graph Learning in Social Internet of Thingscited by 0
  33. 2025A Physics-Augmented GraphGPS Framework for the Reconstruction of 3D Riemann Problems from Sparse Datacited by 0
  34. 2025iN2V: Bringing Transductive Node Embeddings to Inductive Graphscited by 0
  35. 2025How do Probabilistic Graphical Models and Graph Neural Networks Look at Network Data?cited by 0
  36. 2025Divide-Then-Rule: A Cluster-Driven Hierarchical Interpolator for Attribute-Missing Graphscited by 0
  37. 2025On the Impact of Downstream Tasks on Sampling and Reconstructing Noisy Graph Signalscited by 0
  38. 2025Multi-View Graph Feature Propagation for Privacy Preservation and Feature Sparsitycited by 0

50 most cited of 63 All 63 (JSON)