VCG-Bench: Towards A Unified Visual-Centric Benchmark for Structured Generation and Editing

Xiaoyan Su, Peijie Dong, Zhenheng Tang, Song Tang, Yuyao Zhai, Kaitao Lin, Liang Chen, Gai Yuhang, Yuyu Luo, Qiang Wang, Xiaowen Chu
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:115957-115994, 2026.

Abstract

Despite the rapid advancements in Vision-Language Models (VLMs), a critical gap remains in their ability to handle structured, controllable diagrammatic tasks essential for professional workflows, as existing methods predominantly rely on pixel-based synthesis which operates in probabilistic pixel spaces and is inherently limited in editability and fidelity; instead, we propose a new "Diagram-as-Code" paradigm with symbolic logic that leverages mxGraph Extensible Markup Language (XML) for precise diagram generation and editing, and we present VCG-Bench, a unified benchmark for visual-centric mxGraph tasks comprising (1) a taxonomized dataset of 1,449 diverse diagrams spanning 6 domains and 15 sub-domains, (2) a paradigm definition that integrates Generation (Vision-to-Code) and Editability (Code-to-Code), and (3) a Tailored Evaluation Protocol employing multi-dimensional metrics such as mxGraph Execution Success Rate and Style Consistency Score (SCS), where experimental results highlight the challenges faced by current State-of-the-Art (SOTA) VLMs in structured fidelity and instruction compliance, reflecting their vision and reasoning capabilities.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-su26e, title = {{VCG}-Bench: Towards A Unified Visual-Centric Benchmark for Structured Generation and Editing}, author = {Su, Xiaoyan and Dong, Peijie and Tang, Zhenheng and Tang, Song and Zhai, Yuyao and Lin, Kaitao and Chen, Liang and Yuhang, Gai and Luo, Yuyu and Wang, Qiang and Chu, Xiaowen}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {115957--115994}, 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/su26e/su26e.pdf}, url = {https://proceedings.mlr.press/v306/su26e.html}, abstract = {Despite the rapid advancements in Vision-Language Models (VLMs), a critical gap remains in their ability to handle structured, controllable diagrammatic tasks essential for professional workflows, as existing methods predominantly rely on pixel-based synthesis which operates in probabilistic pixel spaces and is inherently limited in editability and fidelity; instead, we propose a new "Diagram-as-Code" paradigm with symbolic logic that leverages mxGraph Extensible Markup Language (XML) for precise diagram generation and editing, and we present VCG-Bench, a unified benchmark for visual-centric mxGraph tasks comprising (1) a taxonomized dataset of 1,449 diverse diagrams spanning 6 domains and 15 sub-domains, (2) a paradigm definition that integrates Generation (Vision-to-Code) and Editability (Code-to-Code), and (3) a Tailored Evaluation Protocol employing multi-dimensional metrics such as mxGraph Execution Success Rate and Style Consistency Score (SCS), where experimental results highlight the challenges faced by current State-of-the-Art (SOTA) VLMs in structured fidelity and instruction compliance, reflecting their vision and reasoning capabilities.} }
Endnote
%0 Conference Paper %T VCG-Bench: Towards A Unified Visual-Centric Benchmark for Structured Generation and Editing %A Xiaoyan Su %A Peijie Dong %A Zhenheng Tang %A Song Tang %A Yuyao Zhai %A Kaitao Lin %A Liang Chen %A Gai Yuhang %A Yuyu Luo %A Qiang Wang %A Xiaowen Chu %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-su26e %I PMLR %P 115957--115994 %U https://proceedings.mlr.press/v306/su26e.html %V 306 %X Despite the rapid advancements in Vision-Language Models (VLMs), a critical gap remains in their ability to handle structured, controllable diagrammatic tasks essential for professional workflows, as existing methods predominantly rely on pixel-based synthesis which operates in probabilistic pixel spaces and is inherently limited in editability and fidelity; instead, we propose a new "Diagram-as-Code" paradigm with symbolic logic that leverages mxGraph Extensible Markup Language (XML) for precise diagram generation and editing, and we present VCG-Bench, a unified benchmark for visual-centric mxGraph tasks comprising (1) a taxonomized dataset of 1,449 diverse diagrams spanning 6 domains and 15 sub-domains, (2) a paradigm definition that integrates Generation (Vision-to-Code) and Editability (Code-to-Code), and (3) a Tailored Evaluation Protocol employing multi-dimensional metrics such as mxGraph Execution Success Rate and Style Consistency Score (SCS), where experimental results highlight the challenges faced by current State-of-the-Art (SOTA) VLMs in structured fidelity and instruction compliance, reflecting their vision and reasoning capabilities.
APA
Su, X., Dong, P., Tang, Z., Tang, S., Zhai, Y., Lin, K., Chen, L., Yuhang, G., Luo, Y., Wang, Q. & Chu, X.. (2026). VCG-Bench: Towards A Unified Visual-Centric Benchmark for Structured Generation and Editing. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:115957-115994 Available from https://proceedings.mlr.press/v306/su26e.html.

Related Material