# Rethinking GNNs and Missing Features: Challenges, Evaluation and a Robust Solution

arXiv: https://arxiv.org/abs/2601.04855 · HTML: https://mlsanity.com/paper/2601.04855/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 52 references that are in the map.
- First posted on arXiv 8 Jan 2026.

## Builds on (most cited first, 26 of 26)

- [XGBoost: A Scalable Tree Boosting System](https://mlsanity.com/paper/1603.02754/) (2016)
- [Open Graph Benchmark: Datasets for Machine Learning on Graphs](https://mlsanity.com/paper/2005.00687/) (2020)
- [Neural Collaborative Filtering](https://mlsanity.com/paper/1708.05031/) (2017)
- [Revisiting Semi-Supervised Learning with Graph Embeddings](https://mlsanity.com/paper/1603.08861/) (2016)
- [BRITS: Bidirectional Recurrent Imputation for Time Series](https://mlsanity.com/paper/1805.10572/) (2018)
- [A Fair Comparison of Graph Neural Networks for Graph Classification](https://mlsanity.com/paper/1912.09893/) (2019)
- [Linear Algebraic Structure of Word Senses, with Applications to Polysemy](https://mlsanity.com/paper/1601.03764/) (2016)
- [Handling Missing Data with Graph Representation Learning](https://mlsanity.com/paper/2010.16418/) (2020)
- [RelBench: A Benchmark for Deep Learning on Relational Databases](https://mlsanity.com/paper/2407.20060/) (2024)
- [Filling the G_ap_s: Multivariate Time Series Imputation by Graph Neural Networks](https://mlsanity.com/paper/2108.00298/) (2021)
- [On the Unreasonable Effectiveness of Feature propagation in Learning on Graphs with Missing Node Features](https://mlsanity.com/paper/2111.12128/) (2021)
- [Graphical Models for Processing Missing Data](https://mlsanity.com/paper/1801.03583/) (2018)
- [Position: Graph Learning Will Lose Relevance Due To Poor Benchmarks](https://mlsanity.com/paper/2502.14546/) (2025)
- [Graph Convolutional Networks for Graphs Containing Missing Features](https://mlsanity.com/paper/2007.04583/) (2020)
- [Fair Attribute Completion on Graph with Missing Attributes](https://mlsanity.com/paper/2302.12977/) (2023)
- [GraphLand: Evaluating Graph Machine Learning Models on Diverse Industrial Data](https://mlsanity.com/paper/2409.14500/) (2024)
- [The Missing Indicator Method: From Low to High Dimensions](https://mlsanity.com/paper/2211.09259/) (2022)
- [No Metric to Rule Them All: Toward Principled Evaluations of Graph-Learning Datasets](https://mlsanity.com/paper/2502.02379/) (2025)
- [Confidence-Based Feature Imputation for Graphs with Partially Known Features](https://mlsanity.com/paper/2305.16618/) (2023)
- [Meta-Path Learning for Multi-relational Graph Neural Networks](https://mlsanity.com/paper/2309.17113/) (2023)
- [GraphBench: Next-generation graph learning benchmarking](https://mlsanity.com/paper/2512.04475/) (2025)
- [Multi-hop Convolutions on Weighted Graphs](https://mlsanity.com/paper/1911.04978/) (2019)
- [Tractable Probabilistic Graph Representation Learning with Graph-Induced Sum-Product Networks](https://mlsanity.com/paper/2305.10544/) (2023)
- [A Self-Explainable Heterogeneous GNN for Relational Deep Learning](https://mlsanity.com/paper/2412.00521/) (2024)
- [Bridging Theory and Practice in Link Representation with Graph Neural Networks](https://mlsanity.com/paper/2506.24018/) (2025)
- [Correcting for non-ignorable missingness in smoking trends](https://mlsanity.com/paper/1502.03609/) (2015)

## Built on by (most cited first, 2 of 2)

- [Gated Graph Attention Networks with Learnable Temperature](https://mlsanity.com/paper/2605.29803/) (2026)
- [When Missingness Follows the Graph: Beyond Feature Propagation for Missing Features](https://mlsanity.com/paper/2610.18113/) (2026)

## Abstract (arXiv)

Handling missing node features is a key challenge for deploying Graph Neural Networks (GNNs) in real-world domains such as healthcare and sensor networks. Existing studies mostly address relatively benign scenarios, namely benchmark datasets with (a) high-dimensional but sparse node features and (b) incomplete data generated under Missing Completely At Random (MCAR) mechanisms. For (a), we theoretically prove that high sparsity substantially limits the information loss caused by missingness, making all models appear robust and preventing a meaningful comparison of their performance. To overcome this limitation, we introduce one synthetic and three real-world datasets with dense, semantically meaningful features. For (b), we move beyond MCAR and design evaluation protocols with more realistic missingness mechanisms. Moreover, we provide a theoretical background to state explicit assumptions on the missingness process and analyze their implications for different methods. Building on this analysis, we propose GNNmim, a simple yet effective baseline for node classification with incomplete feature data. Experiments show that GNNmim is competitive with respect to specialized architectures across diverse datasets and missingness regimes.