Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs

Langzhang Liang, Sunwoo Kim, Kijung Shin, Zenglin Xu, Shirui Pan, Yuan Qi

Abstract

Graph Neural Networks (GNNs) have gained significant attention as a powerful modeling and inference method, especially for homophilic graph-structured data. To empower GNNs in heterophilic graphs, where adjacent nodes exhibit dissimilar labels or features, Signed Message Passing (SMP) has been widely adopted. However, there is a lack of theoretical and empirical analysis regarding the limitations of SMP. In this work, we unveil some potential pitfalls of SMP and their remedies. We first identify two limitations of SMP: undesirable representation update for multi-hop neighbors and vulnerability against oversmoothing issues. To overcome these challenges, we propose a novel message passing function called Multiset to Multiset GNN(M2M-GNN). Our theoretical analyses and extensive experiments demonstrate that M2M-GNN effectively alleviates the aforementioned limitations of SMP, yielding superior performance in comparison

Benchmark Results

Tested on 6 datasets 28 baseline values reported too low

Tables show presented results in global context (best 3 papers + GCN + MLP).

Heterophilic Graphs

Texas

Global leaderboard →

#	Model	Accuracy ↑	Paper
1	P^2GNN Hyb global	0.9672 ±0.0147	$P^2$GNN: Two Prototype Sets to boost GNN Performance
2	DoG code ↗ ext data GNN global	0.9660 ±0.0210	Diffusion on Graph: Augmentation of Graph Structure for Node Classification code ↗
3	GPF-plus code ↗ ext data GNN global	0.9583 ±0.0419	ProG: A Graph Prompt Learning Benchmark code ↗
5	ACM-GCN GNN baseline actual #5 vs paper #91 86 places worse	0.9557 ±0.0360 best reported	actual #5 vs paper #91 86 places worse paper 0.8784 ±0.0440 best at pub: From Node Interaction to Hop Interaction: New Effective and Scalable Graph Learning Paradigm (2022) best today: $P^2$GNN: Two Prototype Sets to boost GNN Performance (2026) paper 0.8784 ±0.0440 Δ today 0.0773
30	GPR-GNN GNN baseline actual #30 vs paper #269 239 places worse	0.9291 ±0.0132 best reported	actual #30 vs paper #269 239 places worse paper 0.7838 ±0.0430 best at pub: Convolutional Neural Networks on Graphs with Chebyshev Approximation, Revisited (NeurIPS 2022) paper 0.7838 ±0.0430 Δ at pub 0.1453
34	MLP Trad baseline actual #34 vs paper #230 196 places worse	0.9230 ±0.0070 best reported	actual #34 vs paper #230 196 places worse paper 0.8081 ±0.0470 best at pub: Design your own universe: a physics-informed agnostic method for enhancing graph neural networks (2024) paper 0.8081 ±0.0470 Δ at pub 0.1149
39	GGCN GNN baseline actual #39 vs paper #148 109 places worse	0.9213 ±0.0305 best reported	actual #39 vs paper #148 109 places worse paper 0.8486 ±0.0450 best at pub: Understanding convolution on graphs via energies (TMLR 2022) best today: Graph Wave Networks (WWW 2025) paper 0.8486 ±0.0450 Δ today 0.0727
40	H2GCN GNN baseline actual #40 vs paper #148 108 places worse	0.9189 ±0.0393 best reported	actual #40 vs paper #148 108 places worse paper 0.8486 ±0.0720 best at pub: LOGIN: A Large Language Model Consulted Graph Neural Network Training Framework (2024) best today: GrokFormer: Graph Fourier Kolmogorov-Arnold Transformers (ICML 2024) Difference within reported uncertainty; not counted as underperformance. paper 0.8486 ±0.0720 Δ at pub 0.0181 Δ today 0.0703
53	GraphSAGE code ↗ GNN global	0.9000 ±0.0365	Graph Neural Networks Need Cluster-Normalize-Activate Modules code ↗
61	FAGCN GNN baseline actual #61 vs paper #289 228 places worse	0.8961 ±0.0152 best reported	actual #61 vs paper #289 228 places worse paper 0.7700 ±0.0770 best at pub: Node-oriented Spectral Filtering for Graph Neural Networks (2022) paper 0.7700 ±0.0770 Δ at pub 0.1261
69	M2M-GNN GNN ★ proposed	0.8919 ±0.0450	This paper
68	DMP GNN baseline actual #68 vs paper #431 363 places worse	0.8919 best reported	actual #68 vs paper #431 363 places worse paper 0.6608 ±0.0700 best at pub: Heterophily-Aware Graph Attention Network (2023) best today: Graph Neural Reaction Diffusion Models Submitted to the editors June 2023. Accepted in March 2024. (2024) paper 0.6608 ±0.0700 Δ at pub 0.1556 Δ today 0.2311
114	Ord. GNN GNN baseline	0.8622 ±0.0410	This paper
164	AERO-GNN GNN baseline	0.8435 ±0.0520	This paper
175	GOAL GNN baseline actual #175 vs paper #175	0.8362 ±0.0670	actual #175 vs paper #175 paper 0.8362 ±0.0670 prev best: ES-GNN: Generalizing Graph Neural Networks Beyond Homophily with Edge Splitting (2022) Improvement within reported uncertainty; not highlighted. paper 0.8362 ±0.0670 Δ at pub 0.1162
211	GAT GNN baseline actual #211 vs paper #576 365 places worse	0.8162 ±0.0645 best reported	actual #211 vs paper #576 365 places worse paper 0.5216 ±0.0660 best at pub: A Quasi-Wasserstein Loss for Learning Graph Neural Networks (WWW 2023) best today: Equivariance Everywhere All At Once: A Recipe for Graph Foundation Models (2025) paper 0.5216 ±0.0660 Δ at pub 0.2784 Δ today 0.2946
257	GCN GNN baseline actual #257 vs paper #568 311 places worse	0.7933 ±0.0447 best reported	actual #257 vs paper #568 311 places worse paper 0.5514 ±0.0510 best at pub: Restructuring Graphs for Higher Homophily via Adaptive Spectral Clustering (AAAI 2022) best today: Graph Wave Networks (WWW 2025) paper 0.5514 ±0.0510 Δ at pub 0.2326 Δ today 0.2419

Cornell

Global leaderboard →

#	Model	Accuracy ↑	Paper
1	P^2GNN Hyb global	0.9541 ±0.0272	$P^2$GNN: Two Prototype Sets to boost GNN Performance
2	ACM-GCN GNN baseline actual #2 vs paper #87 85 places worse	0.9480 ±0.0380 best reported	actual #2 vs paper #87 85 places worse paper 0.8514 ±0.0600 best at pub: From Node Interaction to Hop Interaction: New Effective and Scalable Graph Learning Paradigm (2022) best today: Heterophily-informed Message Passing (TMLR 2025) paper 0.8514 ±0.0600 Δ today 0.0966
3	RF-GCN GNN global	0.9429	Simplifying Graph Convolutional Networks with Redundancy-Free Neighbors
11	GPR-GNN GNN baseline actual #11 vs paper #177 166 places worse	0.9279 ±0.0312 best reported	actual #11 vs paper #177 166 places worse paper 0.8027 ±0.0810 best at pub: Heterophily-Aware Graph Attention Network (2023) paper 0.8027 ±0.0810 Δ at pub 0.1252
15	GGCN GNN baseline actual #15 vs paper #79 64 places worse	0.9230 ±0.0367 best reported	actual #15 vs paper #79 64 places worse paper 0.8568 ±0.0660 best at pub: Understanding convolution on graphs via energies (TMLR 2022) best today: $P^2$GNN: Two Prototype Sets to boost GNN Performance (2026) Difference within reported uncertainty; not counted as underperformance. paper 0.8568 ±0.0660 Δ today 0.0662
24	MLP Trad baseline actual #24 vs paper #158 134 places worse	0.9136 ±0.0070 best reported	actual #24 vs paper #158 134 places worse paper 0.8189 ±0.0640 best at pub: Node-oriented Spectral Filtering for Graph Neural Networks (2022) Difference within reported uncertainty; not counted as underperformance. paper 0.8189 ±0.0640 Δ at pub 0.0947
31	DMP GNN baseline actual #31 vs paper #355 324 places worse	0.8919 best reported	actual #31 vs paper #355 324 places worse paper 0.6273 ±0.0450 best at pub: Heterophily-Aware Graph Attention Network (2023) best today: Graph Neural Reaction Diffusion Models Submitted to the editors June 2023. Accepted in March 2024. (2024) paper 0.6273 ±0.0450 Δ at pub 0.2427 Δ today 0.2646
38	FAGCN GNN baseline actual #38 vs paper #145 107 places worse	0.8852 ±0.0133 best reported	actual #38 vs paper #145 107 places worse paper 0.8241 ±0.0380 best at pub: Node-oriented Spectral Filtering for Graph Neural Networks (2022) paper 0.8241 ±0.0380 Δ at pub 0.0611
53	Ord. GNN GNN baseline	0.8703 ±0.0470	This paper
64	M2M-GNN GNN ★ proposed	0.8648 ±0.0610	This paper
68	H2GCN GNN baseline actual #68 vs paper #136 68 places worse	0.8623 ±0.0471 best reported	actual #68 vs paper #136 68 places worse paper 0.8270 ±0.0520 best at pub: Design your own universe: a physics-informed agnostic method for enhancing graph neural networks (2024) best today: $P^2$GNN: Two Prototype Sets to boost GNN Performance (2026) Difference within reported uncertainty; not counted as underperformance. paper 0.8270 ±0.0520 Δ at pub 0.0240 Δ today 0.0353
81	GOAL GNN baseline actual #81 vs paper #81	0.8568 ±0.0620	actual #81 vs paper #81 paper 0.8568 ±0.0620 prev best: ES-GNN: Generalizing Graph Neural Networks Beyond Homophily with Edge Splitting (2022) paper 0.8568 ±0.0620 Δ at pub 0.3078
98	AERO-GNN GNN baseline actual #98 vs paper #98	0.8480 ±0.0330	actual #98 vs paper #98 paper 0.8480 ±0.0330 prev best: Breaking the Entanglement of Homophily and Heterophily in Semi-supervised Node Classification (2023) Improvement within reported uncertainty; not highlighted. paper 0.8480 ±0.0330 Δ at pub 0.0440
138	SAGE GNN global	0.8269	Language Models as Messengers: Enhancing Message Passing in Heterophilic Graph Learning
158	GAT GNN baseline actual #158 vs paper #358 200 places worse	0.8190 ±0.0540 best reported	actual #158 vs paper #358 200 places worse paper 0.6189 ±0.0500 best at pub: Restructuring Graphs for Higher Homophily via Adaptive Spectral Clustering (AAAI 2022) paper 0.6189 ±0.0500 Δ at pub 0.2001
170	GCN GNN baseline actual #170 vs paper #372 202 places worse	0.8082 ±0.0360 best reported	actual #170 vs paper #372 202 places worse paper 0.6054 ±0.0530 best at pub: Differentiable Cluster Graph Neural Network (2024) best today: Bayesian Neighborhood Adaptation for Graph Neural Networks (TMLR 2025) paper 0.6054 ±0.0530 Δ at pub 0.1961 Δ today 0.2028

Actor

Global leaderboard →

#	Model	Accuracy ↑	Paper
1	DiGGR Hyb global	0.4535 ±0.0353	Disentangled Generative Graph Representation Learning
2	TRIGON code ↗ Hyb global	0.4402 ±0.0028	Dynamic Triangulation-Based Graph Rewiring for Graph Neural Networks code ↗
3	DoG code ↗ ext data GNN global	0.4320 ±0.0120	Diffusion on Graph: Augmentation of Graph Structure for Node Classification code ↗
16	ACM-GCN GNN baseline actual #16 vs paper #173 157 places worse	0.4198 ±0.0089 best reported	actual #16 vs paper #173 157 places worse paper 0.3628 ±0.0100 best at pub: ES-GNN: Generalizing Graph Neural Networks Beyond Homophily with Edge Splitting (2022) best today: $P^2$GNN: Two Prototype Sets to boost GNN Performance (2026) paper 0.3628 ±0.0100 Δ at pub 0.0032 Δ today 0.0570
42	MLP Trad baseline actual #42 vs paper #158 116 places worse	0.4026 ±0.0108 best reported	actual #42 vs paper #158 116 places worse paper 0.3653 ±0.0070 best at pub: Heterophily-Aware Graph Attention Network (2023) paper 0.3653 ±0.0070 Δ at pub 0.0373
53	GPR-GNN GNN baseline actual #53 vs paper #247 194 places worse	0.3992 ±0.0067 best reported	actual #53 vs paper #247 194 places worse paper 0.3463 ±0.0120 best at pub: Convolutional Neural Networks on Graphs with Chebyshev Approximation, Revisited (NeurIPS 2022) best today: Piecewise Constant Spectral Graph Neural Network (TMLR 2025) paper 0.3463 ±0.0120 Δ at pub 0.0528 Δ today 0.0529
56	FAGCN GNN baseline actual #56 vs paper #198 142 places worse	0.3970 best reported	actual #56 vs paper #198 142 places worse paper 0.3567 ±0.0090 best at pub: Beyond Low-frequency Information in Graph Convolutional Networks (AAAI 2021) paper 0.3567 ±0.0090 Δ at pub 0.0403
62	GGCN GNN baseline actual #62 vs paper #113 51 places worse	0.3929 ±0.0048 best reported	actual #62 vs paper #113 51 places worse paper 0.3754 ±0.0150 best at pub: Feature Transportation Improves Graph Neural Networks (AAAI 2023) best today: $P^2$GNN: Two Prototype Sets to boost GNN Performance (2026) Difference within reported uncertainty; not counted as underperformance. paper 0.3754 ±0.0150 Δ at pub 0.0027 Δ today 0.0175
68	GraphSAGE code ↗ GNN global	0.3899 ±0.0085	A Quasi-Wasserstein Loss for Learning Graph Neural Networks code ↗
71	H2GCN GNN baseline actual #71 vs paper #198 127 places worse	0.3885 ±0.0117 best reported	actual #71 vs paper #198 127 places worse paper 0.3570 ±0.0100 best at pub: ES-GNN: Generalizing Graph Neural Networks Beyond Homophily with Edge Splitting (2022) best today: $P^2$GNN: Two Prototype Sets to boost GNN Performance (2026) paper 0.3570 ±0.0100 Δ at pub 0.0270 Δ today 0.0315
92	Ord. GNN GNN baseline	0.3799 ±0.0100	This paper
122	DMP GNN baseline actual #122 vs paper #389 267 places worse	0.3726 ±0.0179 best reported	actual #122 vs paper #389 267 places worse paper 0.2830 ±0.0270 best at pub: Heterophily-Aware Graph Attention Network (2023) paper 0.2830 ±0.0270 Δ at pub 0.0896
135	GCN GNN baseline actual #135 vs paper #408 273 places worse	0.3690 best reported	actual #135 vs paper #408 273 places worse paper 0.2732 ±0.0110 best at pub: Beyond Low-frequency Information in Graph Convolutional Networks (AAAI 2021) paper 0.2732 ±0.0110 Δ at pub 0.0958
147	M2M-GNN GNN ★ proposed	0.3672 ±0.0160	This paper
157	AERO-GNN GNN baseline	0.3657 ±0.0110	This paper
163	GOAL GNN baseline actual #163 vs paper #163	0.3646 ±0.0100	actual #163 vs paper #163 paper 0.3646 ±0.0100 prev best: ES-GNN: Generalizing Graph Neural Networks Beyond Homophily with Edge Splitting (2022) Improvement within reported uncertainty; not highlighted. paper 0.3646 ±0.0100 Δ at pub 0.0016
182	GAT GNN baseline actual #182 vs paper #406 224 places worse	0.3605 ±0.0035 best reported	actual #182 vs paper #406 224 places worse paper 0.2744 ±0.0080 best at pub: Hierarchical Graph Transformer with Adaptive Node Sampling (NeurIPS 2022) paper 0.2744 ±0.0080 Δ at pub 0.0861

Wisconsin

Global leaderboard →

#	Model	Accuracy ↑	Paper
1	DMD++ Hyb global	0.9190 ±0.0260	When Graph Neural Networks Meet Dynamic Mode Decomposition
2	F-GREAD-BS code ↗ GNN global	0.9059 ±0.0380	Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND code ↗
3	APPNP GNN global	0.9010 ±0.0350	When Graph Neural Networks Meet Dynamic Mode Decomposition
11	M2M-GNN GNN ★ proposed	0.8901 ±0.0410	This paper
15	ACM-GCN GNN baseline	0.8843 ±0.0320	This paper
20	Ord. GNN GNN baseline	0.8804 ±0.0360	This paper
22	H2GCN GNN baseline actual #22 vs paper #23 1 place worse	0.8790 ±0.0420 best reported	actual #22 vs paper #23 1 place worse paper 0.8765 ±0.0490 best at pub: Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND (ICLR 2024) best today: When Graph Neural Networks Meet Dynamic Mode Decomposition (ICLR 2024) Difference within reported uncertainty; not counted as underperformance. paper 0.8765 ±0.0490 Δ today 0.0025
30	GOAL GNN baseline	0.8698 ±0.0440	This paper
32	GGCN GNN baseline	0.8686 ±0.0320	This paper
37	MLP Trad baseline actual #37 vs paper #42 5 places worse	0.8629 ±0.0450 best reported	actual #37 vs paper #42 5 places worse paper 0.8529 ±0.0330 best at pub: Is Homophily a Necessity for Graph Neural Networks? (ICLR 2021) Difference within reported uncertainty; not counted as underperformance. paper 0.8529 ±0.0330 Δ at pub 0.0100
53	GAT GNN baseline actual #53 vs paper #213 160 places worse	0.8430 ±0.0370 best reported	actual #53 vs paper #213 160 places worse paper 0.4941 ±0.0400 best at pub: Restructuring Graphs for Higher Homophily via Adaptive Spectral Clustering (AAAI 2022) paper 0.4941 ±0.0400 Δ at pub 0.3489
56	GPR-GNN GNN baseline actual #56 vs paper #64 8 places worse	0.8412 best reported	actual #56 vs paper #64 8 places worse paper 0.8294 ±0.0420 best at pub: Beyond Smoothing: Unsupervised Graph Representation Learning with Edge Heterophily Discriminating (AAAI 2022) Difference within reported uncertainty; not counted as underperformance. paper 0.8294 ±0.0420 Δ at pub 0.0118
60	FAGCN GNN baseline actual #60 vs paper #89 29 places worse	0.8333 best reported	actual #60 vs paper #89 29 places worse paper 0.7832 ±0.0630 best at pub: Beyond Smoothing: Unsupervised Graph Representation Learning with Edge Heterophily Discriminating (AAAI 2022) Difference within reported uncertainty; not counted as underperformance. paper 0.7832 ±0.0630 Δ at pub 0.0501
63	GCN GNN baseline actual #63 vs paper #205 142 places worse	0.8310 ±0.0320 best reported	actual #63 vs paper #205 142 places worse paper 0.5176 ±0.0300 best at pub: Restructuring Graphs for Higher Homophily via Adaptive Spectral Clustering (AAAI 2022) paper 0.5176 ±0.0300 Δ at pub 0.3134
68	AERO-GNN GNN baseline	0.8124 ±0.0680	This paper
69	GraphSAGE code ↗ GNN global	0.8118 ±0.0556	Graph Navier–Stokes Networks code ↗
183	DMP GNN baseline	0.5641 ±0.0550	This paper

Heterophily Benchmark

Amazon-ratings

Global leaderboard →

#	Model	Accuracy ↑	Paper
1	UniGAP global	0.6375 ±0.0110	UniGAP: A Universal and Adaptive Graph Upsampling Approach to Mitigate Over-Smoothing in Node …
2	H2GCN + HalfHop GNN global	0.5961 ±0.0100	UniGAP: A Universal and Adaptive Graph Upsampling Approach to Mitigate Over-Smoothing in Node …
3	GraphSAGE + HalfHop GNN global	0.5897 ±0.0150	UniGAP: A Universal and Adaptive Graph Upsampling Approach to Mitigate Over-Smoothing in Node …
7	H2GCN GNN baseline actual #7 vs paper #196 189 places worse	0.5681 ±0.0090 best reported	actual #7 vs paper #196 189 places worse paper 0.3647 ±0.0020 best at pub: Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND (ICLR 2024) best today: UniGAP: A Universal and Adaptive Graph Upsampling Approach to Mitigate Over-Smoothing in Node Classification Tasks (2024) paper 0.3647 ±0.0020 Δ at pub 0.0489 Δ today 0.2034
12	SAGE code ↗ GNN global	0.5569 ±0.0023	Graph Transductive Sharpening: Leveraging Unlabeled Predictions in Node Classification code ↗
14	GAT GNN baseline actual #14 vs paper #105 91 places worse	0.5559 ±0.0027 best reported	actual #14 vs paper #105 91 places worse paper 0.4909 ±0.0060 best at pub: Graph Neural Networks for Graphs with Heterophily: A Survey (2022) best today: Graph Transductive Sharpening: Leveraging Unlabeled Predictions in Node Classification (2026) paper 0.4909 ±0.0060 Δ today 0.0650
24	GCN GNN baseline actual #24 vs paper #111 87 places worse	0.5411 ±0.0034 best reported	actual #24 vs paper #111 87 places worse paper 0.4870 ±0.0060 best at pub: Graph Neural Networks for Graphs with Heterophily: A Survey (2022) best today: Graph Transductive Sharpening: Leveraging Unlabeled Predictions in Node Classification (2026) paper 0.4870 ±0.0060 Δ today 0.0541
47	ACM-GCN GNN baseline actual #47 vs paper #189 142 places worse	0.5294 ±0.0023 best reported	actual #47 vs paper #189 142 places worse paper 0.3862 ±0.0060 best at pub: Graph Neural Networks for Graphs with Heterophily: A Survey (2022) paper 0.3862 ±0.0060 Δ at pub 0.1432
84	MLP Trad baseline actual #84 vs paper #169 85 places worse	0.5058 best reported	actual #84 vs paper #169 85 places worse paper 0.4287 ±0.0040 best at pub: Understanding Heterophily for Graph Neural Networks (ICML 2024) paper 0.4287 ±0.0040 Δ at pub 0.0771
103	GPR-GNN GNN baseline actual #103 vs paper #148 45 places worse	0.4937 ±0.0071 best reported	actual #103 vs paper #148 45 places worse paper 0.4488 ±0.0030 best at pub: Evolving Computation Graphs (ICML 2023) best today: Piecewise Constant Spectral Graph Neural Network (TMLR 2025) paper 0.4488 ±0.0030 Δ today 0.0449
105	M2M-GNN GNN ★ proposed	0.4918 ±0.0060	This paper
137	FAGCN GNN baseline actual #137 vs paper #156 19 places worse	0.4632 ±0.0250 best reported	actual #137 vs paper #156 19 places worse paper 0.4412 ±0.0030 best at pub: Unleashing the Potential of Fractional Calculus in Graph Neural Networks with FROND (ICLR 2024) best today: Distributed-Order Fractional Graph Operating Network (NeurIPS 2024) Difference within reported uncertainty; not counted as underperformance. paper 0.4412 ±0.0030 Δ at pub 0.0220 Δ today 0.0220
142	AERO-GNN GNN baseline	0.4571 ±0.0050	This paper
165	GGCN GNN baseline actual #165 vs paper #196 31 places worse	0.4300 ±0.0032 best reported	actual #165 vs paper #196 31 places worse paper 0.3686 ±0.0040 best at pub: Polynormer: Polynomial-Expressive Graph Transformer in Linear Time (ICLR 2024) paper 0.3686 ±0.0040 Δ at pub 0.0614
191	Ord. GNN GNN baseline	0.3852 ±0.0040	This paper
195	GOAL GNN baseline	0.3794 ±0.0030	This paper
198	DMP GNN baseline	0.3584 ±0.0040	This paper

LINKX Benchmarks

Penn94

Global leaderboard →

#	Model	Accuracy ↑	Paper
1	DC-GNN GNN global	0.8669 ±0.0022	Differentiable Cluster Graph Neural Network
2	M2M-GNN GNN ★ proposed	0.8594 ±0.0040	This paper
3	GloGNN GNN global	0.8574 ±0.0042	Differentiable Cluster Graph Neural Network
15	GOAL GNN baseline	0.8418 ±0.0030	This paper
16	Ord. GNN GNN baseline	0.8365 ±0.0060	This paper
20	GPR-GNN GNN baseline actual #20 vs paper #37 17 places worse	0.8354 ±0.0032 best reported	actual #20 vs paper #37 17 places worse paper 0.8138 ±0.0010 best at pub: Convolutional Neural Networks on Graphs with Chebyshev Approximation, Revisited (NeurIPS 2022) best today: Optimizing Polynomial Graph Filters: A Novel Adaptive Krylov Subspace Approach (WWW 2024) paper 0.8138 ±0.0010 Δ at pub 0.0216 Δ today 0.0216
24	H2GCN GNN baseline actual #24 vs paper #37 13 places worse	0.8320 ±0.0055 best reported	actual #24 vs paper #37 13 places worse paper 0.8131 ±0.0060 best at pub: Clarify Confused Nodes via Separated Learning (2023) paper 0.8131 ±0.0060 Δ at pub 0.0189
26	ACM-GCN GNN baseline actual #26 vs paper #30 4 places worse	0.8298 ±0.0063 best reported	actual #26 vs paper #30 4 places worse paper 0.8252 ±0.0090 best at pub: Clarify Confused Nodes via Separated Learning (2023) Difference within reported uncertainty; not counted as underperformance. paper 0.8252 ±0.0090 Δ at pub 0.0046
29	GIN code ↗ GNN global	0.8268 ±0.0032	Mixture of Weak & Strong Experts on Graphs code ↗
31	GCN GNN baseline	0.8247 ±0.0020	This paper
32	AERO-GNN GNN baseline	0.8247 ±0.0070	This paper
36	GAT GNN baseline	0.8153 ±0.0050	This paper
46	FAGCN GNN baseline actual #46 vs paper #58 12 places worse	0.7697 ±0.0069 best reported	actual #46 vs paper #58 12 places worse paper 0.7285 ±0.0050 best at pub: Clarify Confused Nodes via Separated Learning (2023) paper 0.7285 ±0.0050 Δ at pub 0.0412
51	MLP Trad baseline actual #51 vs paper #57 6 places worse	0.7534 ±0.0064 best reported	actual #51 vs paper #57 6 places worse paper 0.7361 ±0.0040 best at pub: Clarify Confused Nodes via Separated Learning (2023) paper 0.7361 ±0.0040 Δ at pub 0.0173
56	DMP GNN baseline	0.7385 ±0.0070	This paper

SOTA performance (not significantly different from the best) Proposed method (this paper) Baseline / comparison (this paper) Global context (other papers) Best at pub — reported result worse than the best known when published (best reported in green) Best at pub — reported value improved over the best known when published Best today — OK at publication; today's leaderboard best is elsewhere