OGB
ogbn-products
Structure
1Graphs
2,449,029Nodes
61,859,140Edges
47Classes
100Node Feature Dim
0.81Homophily
Coverage
535Results
1Variants
83Papers
- ICML 2026: Clustering as Reasoning: A $k$-Means Interpretation of Chain-of-Thought Graph Learning Clustering as Reasoning: A $k$-Means Interpretation of Chain-of-Thought Graph Learning ICML 2026
- ATLAS: Adaptive Topology-based Learning at Scale for Homophilic and Heterophilic Graphs ATLAS: Adaptive Topology-based Learning at Scale for Homophilic and Heterophilic Graphs
- CORE: Contrastive Masked Feature Reconstruction on Graphs CORE: Contrastive Masked Feature Reconstruction on Graphs
- Quantum Graph Attention Network: A Novel Quantum Multi-Head Attention Mechanism for Graph Learning Quantum Graph Attention Network: A Novel Quantum Multi-Head Attention Mechanism for Graph Learning
- Hypergraph Neural Network with State Space Models for Node Classification Hypergraph Neural Network with State Space Models for Node Classification
- KDD 2025: Quantizing Text-attributed Graphs for Semantic-Structural Integration Quantizing Text-attributed Graphs for Semantic-Structural Integration KDD 2025
- LPS-GNN: Deploying Graph Neural Networks on Graphs with 100-Billion Edges LPS-GNN: Deploying Graph Neural Networks on Graphs with 100-Billion Edges
- TMLR 2025: Efficient Text-Attributed Graph Learning through Selective Annotation and Graph Alignment Efficient Text-Attributed Graph Learning through Selective Annotation and Graph Alignment TMLR 2025
- Language Models as Messengers: Enhancing Message Passing in Heterophilic Graph Learning Language Models as Messengers: Enhancing Message Passing in Heterophilic Graph Learning
- Efficient Mixed Precision Quantization in Graph Neural Networks Efficient Mixed Precision Quantization in Graph Neural Networks
- Federated Prototype Graph Learning Federated Prototype Graph Learning
- ICLR 2025: Accurate and Scalable Graph Neural Networks via Message Invariance Accurate and Scalable Graph Neural Networks via Message Invariance ICLR 2025
- Democratizing Large Language Model-Based Graph Data Augmentation via Latent Knowledge Graphs Democratizing Large Language Model-Based Graph Data Augmentation via Latent Knowledge Graphs
- NeurIPS 2025: A Signed Graph Approach to Understanding and Mitigating Oversmoothing in GNNs A Signed Graph Approach to Understanding and Mitigating Oversmoothing in GNNs NeurIPS 2025
- GraphICL: Unlocking Graph Learning Potential in LLMs through Structured Prompt Design GraphICL: Unlocking Graph Learning Potential in LLMs through Structured Prompt Design
- ICML 2024: Towards Graph Foundation Models: Learning Generalities Across Graphs via Task-Trees Towards Graph Foundation Models: Learning Generalities Across Graphs via Task-Trees ICML 2024
- TMLR 2024: Sparse Decomposition of Graph Neural Networks Sparse Decomposition of Graph Neural Networks TMLR 2024
- IGNN-Solver: A Graph Neural Solver for Implicit Graph Neural Networks IGNN-Solver: A Graph Neural Solver for Implicit Graph Neural Networks
- How to Make LMs Strong Node Classifiers? How to Make LMs Strong Node Classifiers?
- All Against Some: Efficient Integration of Large Language Models for Message Passing in Graph Neural Networks All Against Some: Efficient Integration of Large Language Models for Message Passing in Graph Neural Networks
- UniGLM: Training One Unified Language Model for Text-Attributed Graph Embedding UniGLM: Training One Unified Language Model for Text-Attributed Graph Embedding
- NeurIPS 2024: Text-space Graph Foundation Models: Comprehensive Benchmarks and New Insights Text-space Graph Foundation Models: Comprehensive Benchmarks and New Insights NeurIPS 2024
- Automatic Graph Topology-Aware Transformer Automatic Graph Topology-Aware Transformer
- ICLR 2024: Fully-inductive Node Classification on Arbitrary Graphs Fully-inductive Node Classification on Arbitrary Graphs ICLR 2024
- NeurIPS 2024: Spatio-Spectral Graph Neural Networks Spatio-Spectral Graph Neural Networks NeurIPS 2024
- ICLR 2024: Node Identifiers: Compact, Discrete Representations for Efficient Graph Learning Node Identifiers: Compact, Discrete Representations for Efficient Graph Learning ICLR 2024
- ICLR 2024: Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND ICLR 2024
- Simple Graph Condensation Simple Graph Condensation
- ICML 2024: Unifews: You Need Fewer Operations for Efficient Graph Neural Networks Unifews: You Need Fewer Operations for Efficient Graph Neural Networks ICML 2024
- CDC: A Simple Framework for Complex Data Clustering CDC: A Simple Framework for Complex Data Clustering
- Smoothed Graph Contrastive Learning via Seamless Proximity Integration Smoothed Graph Contrastive Learning via Seamless Proximity Integration
- KDD 2024: UniGraph: Learning a Unified Cross-Domain Foundation Model for Text-Attributed Graphs UniGraph: Learning a Unified Cross-Domain Foundation Model for Text-Attributed Graphs KDD 2024
- WWW 2024: Rethinking Node-wise Propagation for Large-scale Graph Learning Rethinking Node-wise Propagation for Large-scale Graph Learning WWW 2024
- IJCAI 2024: Efficient Tuning and Inference for Large Language Models on Textual Graphs Efficient Tuning and Inference for Large Language Models on Textual Graphs IJCAI 2024
- WWW 2024: Disentangled Condensation for Large-scale Graphs Disentangled Condensation for Large-scale Graphs WWW 2024
- Graph Transformers for Large Graphs Graph Transformers for Large Graphs
- Rethinking and Simplifying Bootstrapped Graph Latents Rethinking and Simplifying Bootstrapped Graph Latents
- ICLR 2023: Mixture of Weak & Strong Experts on Graphs Mixture of Weak & Strong Experts on Graphs ICLR 2023
- ICLR 2023: MuseGNN: Forming Scalable, Convergent GNN Layers that Minimize a Sampling-Based Energy MuseGNN: Forming Scalable, Convergent GNN Layers that Minimize a Sampling-Based Energy ICLR 2023
- Accelerating Scalable Graph Neural Network Inference with Node-Adaptive Propagation Accelerating Scalable Graph Neural Network Inference with Node-Adaptive Propagation
- ICLR 2023: Label-free Node Classification on Graphs with Large Language Models (LLMs) Label-free Node Classification on Graphs with Large Language Models (LLMs) ICLR 2023
- Label Deconvolution for Node Representation Learning on Large-scale Attributed Graphs against Learning Bias Label Deconvolution for Node Representation Learning on Large-scale Attributed Graphs against Learning Bias
- CaT: Balanced Continual Graph Learning with Graph Condensation CaT: Balanced Continual Graph Learning with Graph Condensation
- FedGTA: Topology-aware Averaging for Federated Graph Learning FedGTA: Topology-aware Averaging for Federated Graph Learning
- SimTeG: A Frustratingly ple Approach Improves xtual raph Learning SimTeG: A Frustratingly ple Approach Improves xtual raph Learning
- ICML 2023: Graph Ladling: Shockingly Simple Parallel GNN Training without Intermediate Communication Graph Ladling: Shockingly Simple Parallel GNN Training without Intermediate Communication ICML 2023
- KDD 2023: Improving Expressivity of GNNs with Subgraph-specific Factor Embedded Normalization Improving Expressivity of GNNs with Subgraph-specific Factor Embedded Normalization KDD 2023
- ICML 2023: Node Embedding from Neural Hamiltonian Orbits in Graph Neural Networks Node Embedding from Neural Hamiltonian Orbits in Graph Neural Networks ICML 2023
- IJCAI 2023: Causal-Based Supervision of Attention in Graph Neural Network: A Better and Simpler Choice towards Powerful Attention Causal-Based Supervision of Attention in Graph Neural Network: A Better and Simpler Choice towards Powerful Attention IJCAI 2023
- IJCAI 2023: Hierarchical Transformer for Scalable Graph Learning Hierarchical Transformer for Scalable Graph Learning IJCAI 2023
- Train Your Own GNN Teacher: Graph-Aware Distillation on Textual Graphs Train Your Own GNN Teacher: Graph-Aware Distillation on Textual Graphs
- LoG 2022: You Can Have Better Graph Neural Networks by Not Training Weights at All: Finding Untrained GNNs Tickets You Can Have Better Graph Neural Networks by Not Training Weights at All: Finding Untrained GNNs Tickets LoG 2022
- ICLR 2022: Latent Graph Inference using Product Manifolds Latent Graph Inference using Product Manifolds ICLR 2022
- From Node Interaction to Hop Interaction: New Effective and Scalable Graph Learning Paradigm From Node Interaction to Hop Interaction: New Effective and Scalable Graph Learning Paradigm
- Efficient Graph Neural Network Inference at Large Scale Efficient Graph Neural Network Inference at Large Scale
- NeurIPS 2022: A Comprehensive Study on Large-Scale Graph Training: Benchmarking and Rethinking A Comprehensive Study on Large-Scale Graph Training: Benchmarking and Rethinking NeurIPS 2022
- NeurIPS 2022: Hierarchical Graph Transformer with Adaptive Node Sampling Hierarchical Graph Transformer with Adaptive Node Sampling NeurIPS 2022
- ICLR 2022: Learning on Large-scale Text-attributed Graphs via Variational Inference Learning on Large-scale Text-attributed Graphs via Variational Inference ICLR 2022
- KDD 2022: Graph Attention Multi-Layer Perceptron Graph Attention Multi-Layer Perceptron KDD 2022
- KDD 2022: Model Degradation Hinders Deep Graph Neural Networks Model Degradation Hinders Deep Graph Neural Networks KDD 2022
- ICML 2022: Alternately Optimized Graph Neural Networks Alternately Optimized Graph Neural Networks ICML 2022
- NeurIPS 2022: Rethinking and Scaling Up Graph Contrastive Learning: An Extremely Efficient Approach with Group Discrimination Rethinking and Scaling Up Graph Contrastive Learning: An Extremely Efficient Approach with Group Discrimination NeurIPS 2022
- KDD 2022: Compressing Deep Graph Neural Networks via Adversarial Knowledge Distillation Compressing Deep Graph Neural Networks via Adversarial Knowledge Distillation KDD 2022
- WWW 2022: PaSca: a Graph Neural Architecture Search System under the Scalable Paradigm PaSca: a Graph Neural Architecture Search System under the Scalable Paradigm WWW 2022
- Neighbor2Seq: Deep Learning on Massive Graphs by Transforming Neighbors to Sequences Neighbor2Seq: Deep Learning on Massive Graphs by Transforming Neighbors to Sequences
- Graph Representation Learning via Aggregation Enhancement Graph Representation Learning via Aggregation Enhancement
- SCR: Training Graph Neural Networks with Consistency Regularization SCR: Training Graph Neural Networks with Consistency Regularization
- On the Unreasonable Effectiveness of Feature Propagation in Learning on Graphs with Missing Node Features On the Unreasonable Effectiveness of Feature Propagation in Learning on Graphs with Missing Node Features
- Sequential Aggregation and Rematerialization: Distributed Full-batch Training of Graph Neural Networks on Large Graphs Sequential Aggregation and Rematerialization: Distributed Full-batch Training of Graph Neural Networks on Large Graphs
- ICLR 2021: Node Feature Extraction by Self-Supervised Multi-scale Neighborhood Prediction Node Feature Extraction by Self-Supervised Multi-scale Neighborhood Prediction ICLR 2021
- CAP: Co-Adversarial Perturbation on Weights and Features for Improving Generalization of Graph Neural Networks CAP: Co-Adversarial Perturbation on Weights and Features for Improving Generalization of Graph Neural Networks
- NeurIPS 2021: On Provable Benefits of Depth in Training Graph Convolutional Networks On Provable Benefits of Depth in Training Graph Convolutional Networks NeurIPS 2021
- Adaptive Label Smoothing To Regularize Large-Scale Graph Training Adaptive Label Smoothing To Regularize Large-Scale Graph Training
- Graph Attention MLP with Reliable Label Utilization Graph Attention MLP with Reliable Label Utilization
- Evaluating Deep Graph Neural Networks Evaluating Deep Graph Neural Networks
- Residual Network and Embedding Usage: New Tricks of Node Classification with Graph Convolutional Networks Residual Network and Embedding Usage: New Tricks of Node Classification with Graph Convolutional Networks
- Scalable and Adaptive Graph Neural Networks with Self-Label-Enhanced Training Scalable and Adaptive Graph Neural Networks with Self-Label-Enhanced Training
- DistGNN: Scalable Distributed Training for Large-Scale Graph Neural Networks DistGNN: Scalable Distributed Training for Large-Scale Graph Neural Networks
- TMLR 2021: Calibrating and Improving Graph Contrastive Learning Calibrating and Improving Graph Contrastive Learning TMLR 2021
- IJCAI 2020: Rethinking the Promotion Brought by Contrastive Learning to Semi-Supervised Node Classification Rethinking the Promotion Brought by Contrastive Learning to Semi-Supervised Node Classification IJCAI 2020
- ICLR 2020: Combining Label Propagation and Simple Models out-performs Graph Neural Networks Combining Label Propagation and Simple Models out-performs Graph Neural Networks ICLR 2020
- IJCAI 2020: Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification IJCAI 2020
- SIGN: Scalable Inception Graph Neural Networks SIGN: Scalable Inception Graph Neural Networks
47-class node classification Predict one of 47 top-level product categories for every product node using its description representation and co-purchase neighborhood. Evaluated by Accuracy.
Dataset details
Task
Predict one of 47 top-level product categories for every product node using its description representation and co-purchase neighborhood. Accuracy is the official metric.
Data
The undirected, unweighted graph contains 2,449,029 Amazon products; an edge indicates that two products are frequently purchased together. Categories and preprocessing follow the Cluster-GCN version of the source collection.
Features
Product descriptions are converted to bag-of-words vectors and reduced by principal component analysis to 100 numeric node features. Sales rank is split metadata, not part of the standard feature vector.
Splits and evaluation
Products are sorted by sales rank: the most popular 8% train the model, the next 2% form validation, and the remaining 90% form test. This deliberately tests transfer from heavily labeled popular items to less popular ones.
Quirks and pitfalls
The split is highly asymmetric and distribution-shifted; a conventional random 90/10 split is a different, easier protocol. The graph contains a very small number of repeated self-loops. Co-purchase edges reflect platform behavior and exposure as well as product similarity.
Sources
SOTA Progression
new SOTA
matches SOTA at that time
other proposed result
other baseline result
Leaderboard
| # | Model | Accuracy ↑ | ACC | AP | F1 | F1-micro | Micro-F1 | Paper |
|---|---|---|---|---|---|---|---|---|
| 1 | Jacobi | 0.8968 | — | — | — | — | — | |
| 2 | GPR | 0.8953 | — | — | — | — | — | |
| 3 | LEMP4HG | 0.8939 | — | — | — | — | — | |
| 4 | SAGE | 0.8933 | — | — | — | — | — | |
| 5 | OGNN | 0.8908 | — | — | — | — | — | |
| 6 | GCN | 0.8868 | — | — | — | — | — | |
| 7 | JKNet | 0.8862 | — | — | — | — | — | |
| 8 | RevGAT | 0.8859 | — | — | — | — | — | |
| 9 | GAT | 0.8829 | — | — | — | — | — | |
| 10 | LLMs | 0.8820
±
0.0005 | — | — | — | — | — | |
| 11 | EnGCN | 0.8799
±
0.0004 | — | — | — | — | — | |
| 12 | Raw Text | 0.8758
±
0.0548 | — | — | — | — | — | |
| 13 | GCNII | 0.8738 | — | — | — | — | — | |
| 14 | GLEM | 0.8736
±
0.0007 | — | — | — | — | — | |
| 15 | SAGN | 0.8736
±
0.0007 | — | — | — | — | — | |
| 16 | GLEM-GNN | 0.8736
±
0.0007 | — | — | — | — | — | |
| 17 | Single | 0.8735 | — | — | — | — | — | |
| 18 | GAT | 0.8722
±
0.0005 | — | — | — | — | — | |
| 19 | SAGN | 0.8718
±
0.0004 | — | — | — | — | — | |
| 20 | GraD-BERT + SAGN+ | 0.8692
±
0.0007 | — | — | — | — | — | |
| 21 | GCN | 0.8691
±
0.0005 | — | — | — | — | — | |
| 22 | Co-train | 0.8684 | — | — | — | — | — | |
| 23 | SCR | 0.8680
±
0.0007 | — | — | — | — | — | |
| 24 | SCR-m | 0.8673
±
0.0008 | — | — | — | — | — | |
| 25 | GIANT-XRT+SAGN+MCR+C&S | 0.8673
±
0.0008 | — | — | — | — | — | |
| 26 | LC | 0.8645
±
0.0013 | — | — | — | — | — | |
| 27 | LD | 0.8645
±
0.0012 | — | — | — | — | — | |
| 28 | SLE | 0.8643
±
0.0020 | — | — | — | — | — | |
| 29 | DGI | 0.8637
±
0.0019 | — | — | — | — | — | |
| 30 | MLP+LLM_features | 0.8634 | — | — | — | — | — | |
| 31 | Cheby | 0.8634 | — | — | — | — | — | |
| 32 | APPNP | 0.8634 | — | — | — | — | — | |
| 33 | GraphMAE | 0.8630
±
0.0007 | — | — | — | — | — | |
| 34 | ML | 0.8628
±
0.0020 | — | — | — | — | — | |
| 35 | OFA | 0.8623
±
0.0007 | — | — | — | — | — | |
| 36 | GMI | 0.8614
±
0.0016 | — | — | — | — | — | |
| 37 | X-GIANT | 0.8612
±
0.0034 | — | — | — | — | — | |
| 38 | mpnet | 0.8560 | — | — | — | — | — | |
| 39 | minilm | 0.8558 | — | — | — | — | — | |
| 40 | OGB-feat + GraD-BERT | 0.8551
±
0.0051 | — | — | — | — | — | |
| 41 | GIANT-XRT | 0.8547
±
0.0029 | — | — | — | — | — | |
| 42 | SimTEG | 0.8540
±
0.0028 | — | — | — | — | — | |
| 43 | RLU+SCR | 0.8520
±
0.0008 | — | — | — | — | — | |
| 44 | C&S | 0.8511
±
0.0007 | — | — | — | — | — | |
| 45 | X_GLEM | 0.8509
±
0.0021 | — | — | — | — | — | |
| 46 | SAGN+SLE | 0.8501
±
0.0010 | — | — | — | — | — | |
| 47 | OGB-feat + SAGN+ | 0.8485
±
0.0010 | — | — | — | — | — | |
| 48 | GLEM-LM | 0.8483
±
0.0004 | — | — | — | — | — | |
| 49 | PaSca-V3 | 0.8480
±
0.0070 | — | — | — | — | — | |
| 50 | ShaDow+ATP | 0.8480
±
0.0020 | — | — | — | — | — |
SOTA performance
set a new SOTA at the time
matched the SOTA at the time