@inproceedings{kakizaki-etal-2026-evoformat,
title = "{E}vo{F}ormat: Discovering Effective Graph-to-Text Formats via Rule-Based Evolutionary Search",
author = "Kakizaki, Kazuya and
Miyagawa, Taiki and
Oiso, Hideyuki and
Furukawa, Ryo",
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.1845/",
doi = "10.18653/2026.findings-emnlp.1845",
pages = "39404--39415",
ISBN = "979-8-89176-462-0",
abstract = "Encoding graphs as text prompts enables LLM-based graph reasoning. However, effective graph-to-text encodings depend on the downstream task, graph structure, and target LLM, making manual format design time-consuming and brittle. We propose EvoFormat, a black-box evolutionary search framework that automatically discovers interpretable graph-to-text formats represented as human-readable formatting rules over node naming, relation expression, and global arrangement. Using only task-level feedback from LLM outputs, EvoFormat iteratively mutates and selects rules without gradients or internal model access. Across synthetic GraphQA and IMDB-MULTI-derived structural QA tasks over unattributed graphs, EvoFormat outperforms strong hand-crafted encodings in most settings."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kakizaki-etal-2026-evoformat">
<titleInfo>
<title>EvoFormat: Discovering Effective Graph-to-Text Formats via Rule-Based Evolutionary Search</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kazuya</namePart>
<namePart type="family">Kakizaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Taiki</namePart>
<namePart type="family">Miyagawa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hideyuki</namePart>
<namePart type="family">Oiso</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ryo</namePart>
<namePart type="family">Furukawa</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>Encoding graphs as text prompts enables LLM-based graph reasoning. However, effective graph-to-text encodings depend on the downstream task, graph structure, and target LLM, making manual format design time-consuming and brittle. We propose EvoFormat, a black-box evolutionary search framework that automatically discovers interpretable graph-to-text formats represented as human-readable formatting rules over node naming, relation expression, and global arrangement. Using only task-level feedback from LLM outputs, EvoFormat iteratively mutates and selects rules without gradients or internal model access. Across synthetic GraphQA and IMDB-MULTI-derived structural QA tasks over unattributed graphs, EvoFormat outperforms strong hand-crafted encodings in most settings.</abstract>
<identifier type="citekey">kakizaki-etal-2026-evoformat</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1845</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1845/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>39404</start>
<end>39415</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T EvoFormat: Discovering Effective Graph-to-Text Formats via Rule-Based Evolutionary Search
%A Kakizaki, Kazuya
%A Miyagawa, Taiki
%A Oiso, Hideyuki
%A Furukawa, Ryo
%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 kakizaki-etal-2026-evoformat
%X Encoding graphs as text prompts enables LLM-based graph reasoning. However, effective graph-to-text encodings depend on the downstream task, graph structure, and target LLM, making manual format design time-consuming and brittle. We propose EvoFormat, a black-box evolutionary search framework that automatically discovers interpretable graph-to-text formats represented as human-readable formatting rules over node naming, relation expression, and global arrangement. Using only task-level feedback from LLM outputs, EvoFormat iteratively mutates and selects rules without gradients or internal model access. Across synthetic GraphQA and IMDB-MULTI-derived structural QA tasks over unattributed graphs, EvoFormat outperforms strong hand-crafted encodings in most settings.
%R 10.18653/2026.findings-emnlp.1845
%U https://aclanthology.org/2026.findings-emnlp.1845/
%U https://doi.org/10.18653/2026.findings-emnlp.1845
%P 39404-39415
Markdown (Informal)
[EvoFormat: Discovering Effective Graph-to-Text Formats via Rule-Based Evolutionary Search](https://aclanthology.org/2026.findings-emnlp.1845/) (Kakizaki et al., Findings 2026)
ACL