@inproceedings{park-etal-2026-benchmark,
title = "A Benchmark for the Generation and Evaluation of Scientific Architecture Diagrams in {HTML}/{SVG}",
author = "Park, Kyeongman and
Lee, Kang-il and
Han, Janghoon and
Hwang, Yerin and
Kang, Taegwan and
Lee, Minwoo and
Lee, Changho and
Jung, Kyomin",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2060/",
pages = "43868--43887",
ISBN = "979-8-89176-462-0",
abstract = "While Large Language Models (LLMs) can now automate many parts of scientific writing, architectural figure generation faces a critical modality bottleneck because image outputs are inherently difficult to edit, verify, and integrate into text-native LLM workflows. To enable fully text-driven synthesis, we propose a new task: generating architectural figures as pure HTML/SVG source code. Furthermore, we introduce FigureCode, a novel benchmark that evaluates structural fidelity and visual properties through code-based evaluation. Our benchmark improves agreement with human judgments by 0.35 Kendall{'}s {\ensuremath{\tau}} over prior VLM-based methods. Moreover, the figure-as-code paradigm improves edit-instruction success rate by 30{\%} and retrieval F1 score by 25.3{\%} over baselines, while achieving 66.1{\%} WCAG/ARIA accessibility coverage."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="park-etal-2026-benchmark">
<titleInfo>
<title>A Benchmark for the Generation and Evaluation of Scientific Architecture Diagrams in HTML/SVG</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kyeongman</namePart>
<namePart type="family">Park</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kang-il</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Janghoon</namePart>
<namePart type="family">Han</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yerin</namePart>
<namePart type="family">Hwang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Taegwan</namePart>
<namePart type="family">Kang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Minwoo</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Changho</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kyomin</namePart>
<namePart type="family">Jung</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>While Large Language Models (LLMs) can now automate many parts of scientific writing, architectural figure generation faces a critical modality bottleneck because image outputs are inherently difficult to edit, verify, and integrate into text-native LLM workflows. To enable fully text-driven synthesis, we propose a new task: generating architectural figures as pure HTML/SVG source code. Furthermore, we introduce FigureCode, a novel benchmark that evaluates structural fidelity and visual properties through code-based evaluation. Our benchmark improves agreement with human judgments by 0.35 Kendall’s \ensuremathτ over prior VLM-based methods. Moreover, the figure-as-code paradigm improves edit-instruction success rate by 30% and retrieval F1 score by 25.3% over baselines, while achieving 66.1% WCAG/ARIA accessibility coverage.</abstract>
<identifier type="citekey">park-etal-2026-benchmark</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2060/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>43868</start>
<end>43887</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Benchmark for the Generation and Evaluation of Scientific Architecture Diagrams in HTML/SVG
%A Park, Kyeongman
%A Lee, Kang-il
%A Han, Janghoon
%A Hwang, Yerin
%A Kang, Taegwan
%A Lee, Minwoo
%A Lee, Changho
%A Jung, Kyomin
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F park-etal-2026-benchmark
%X While Large Language Models (LLMs) can now automate many parts of scientific writing, architectural figure generation faces a critical modality bottleneck because image outputs are inherently difficult to edit, verify, and integrate into text-native LLM workflows. To enable fully text-driven synthesis, we propose a new task: generating architectural figures as pure HTML/SVG source code. Furthermore, we introduce FigureCode, a novel benchmark that evaluates structural fidelity and visual properties through code-based evaluation. Our benchmark improves agreement with human judgments by 0.35 Kendall’s \ensuremathτ over prior VLM-based methods. Moreover, the figure-as-code paradigm improves edit-instruction success rate by 30% and retrieval F1 score by 25.3% over baselines, while achieving 66.1% WCAG/ARIA accessibility coverage.
%U https://aclanthology.org/2026.findings-emnlp.2060/
%P 43868-43887
Markdown (Informal)
[A Benchmark for the Generation and Evaluation of Scientific Architecture Diagrams in HTML/SVG](https://aclanthology.org/2026.findings-emnlp.2060/) (Park et al., Findings 2026)
ACL
- Kyeongman Park, Kang-il Lee, Janghoon Han, Yerin Hwang, Taegwan Kang, Minwoo Lee, Changho Lee, and Kyomin Jung. 2026. A Benchmark for the Generation and Evaluation of Scientific Architecture Diagrams in HTML/SVG. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 43868–43887, Budapest, Hungary. Association for Computational Linguistics.