OpenMAG: A Comprehensive Benchmark for Multimodal-Attributed Graph

Chenxi Wan, Xunkai Li, Yilong Zuo, Haokun Deng, Sihan Li, Bowen Fan, Hongchao Qin, Rong-Hua Li, Guoren Wang
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:124554-124579, 2026.

Abstract

Multimodal-Attributed Graph (MAG) learning has achieved remarkable success in modeling complex real-world systems by integrating graph topology with rich attributes from multiple modalities. With the rapid proliferation of novel MAG models capable of handling intricate cross-modal semantics and structural dependencies, establishing a rigorous and unified evaluation standard has become imperative. Although existing benchmarks have facilitated initial progress, they exhibit critical limitations in domain coverage, encoder flexibility, model diversity, and task scope, presenting significant challenges to fair evaluation. To bridge this gap, we present OpenMAG, a comprehensive benchmark that integrates 19 datasets across 6 domains and incorporates 16 encoders to support both static and trainable feature encoding. OpenMAG further implements a standardized library of 24 state-of-the-art models and supports 8 downstream tasks, enabling fair comparisons within a unified framework. Through systematic assessment of necessity, data quality, effectiveness, robustness, and efficiency, we derive 14 fundamental insights into MAG learning to guide future advancements. Our code is available at https://github.com/YUKI-N810/OpenMAG.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-wan26h, title = {{O}pen{MAG}: A Comprehensive Benchmark for Multimodal-Attributed Graph}, author = {Wan, Chenxi and Li, Xunkai and Zuo, Yilong and Deng, Haokun and Li, Sihan and Fan, Bowen and Qin, Hongchao and Li, Rong-Hua and Wang, Guoren}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {124554--124579}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/wan26h/wan26h.pdf}, url = {https://proceedings.mlr.press/v306/wan26h.html}, abstract = {Multimodal-Attributed Graph (MAG) learning has achieved remarkable success in modeling complex real-world systems by integrating graph topology with rich attributes from multiple modalities. With the rapid proliferation of novel MAG models capable of handling intricate cross-modal semantics and structural dependencies, establishing a rigorous and unified evaluation standard has become imperative. Although existing benchmarks have facilitated initial progress, they exhibit critical limitations in domain coverage, encoder flexibility, model diversity, and task scope, presenting significant challenges to fair evaluation. To bridge this gap, we present OpenMAG, a comprehensive benchmark that integrates 19 datasets across 6 domains and incorporates 16 encoders to support both static and trainable feature encoding. OpenMAG further implements a standardized library of 24 state-of-the-art models and supports 8 downstream tasks, enabling fair comparisons within a unified framework. Through systematic assessment of necessity, data quality, effectiveness, robustness, and efficiency, we derive 14 fundamental insights into MAG learning to guide future advancements. Our code is available at https://github.com/YUKI-N810/OpenMAG.} }
Endnote
%0 Conference Paper %T OpenMAG: A Comprehensive Benchmark for Multimodal-Attributed Graph %A Chenxi Wan %A Xunkai Li %A Yilong Zuo %A Haokun Deng %A Sihan Li %A Bowen Fan %A Hongchao Qin %A Rong-Hua Li %A Guoren Wang %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-wan26h %I PMLR %P 124554--124579 %U https://proceedings.mlr.press/v306/wan26h.html %V 306 %X Multimodal-Attributed Graph (MAG) learning has achieved remarkable success in modeling complex real-world systems by integrating graph topology with rich attributes from multiple modalities. With the rapid proliferation of novel MAG models capable of handling intricate cross-modal semantics and structural dependencies, establishing a rigorous and unified evaluation standard has become imperative. Although existing benchmarks have facilitated initial progress, they exhibit critical limitations in domain coverage, encoder flexibility, model diversity, and task scope, presenting significant challenges to fair evaluation. To bridge this gap, we present OpenMAG, a comprehensive benchmark that integrates 19 datasets across 6 domains and incorporates 16 encoders to support both static and trainable feature encoding. OpenMAG further implements a standardized library of 24 state-of-the-art models and supports 8 downstream tasks, enabling fair comparisons within a unified framework. Through systematic assessment of necessity, data quality, effectiveness, robustness, and efficiency, we derive 14 fundamental insights into MAG learning to guide future advancements. Our code is available at https://github.com/YUKI-N810/OpenMAG.
APA
Wan, C., Li, X., Zuo, Y., Deng, H., Li, S., Fan, B., Qin, H., Li, R. & Wang, G.. (2026). OpenMAG: A Comprehensive Benchmark for Multimodal-Attributed Graph. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:124554-124579 Available from https://proceedings.mlr.press/v306/wan26h.html.

Related Material