@inproceedings{wang-etal-2026-deeppersona,
title = "{D}eep{P}ersona: Scaling Persona Depth via Taxonomy-Guided Generation",
author = "Wang, Zhen and
Zhou, Yufan and
Luo, Zhongyan and
Ye, Lyumanshan and
Wood, Adam and
Mansour, Saab and
Yao, Man and
Pan, Luoshang",
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.2229/",
pages = "47721--47742",
ISBN = "979-8-89176-462-0",
abstract = "Synthetic personas support LLM personalization, social simulation, and alignment evaluation, yet existing resources typically represent each individual with only a small number of structured attributes. We introduce DeepPersona, a generative engine that scales persona depth to roughly 200 instantiated taxonomy nodes while preserving narrative coherence. DeepPersona combines (i) a hierarchical taxonomy of 8,496 human-descriptive attributes mined from 62k personalizable QA pairs in real user-chatbot dialogues and (ii) a progressive sampler that conditions each new attribute on the evolving persona and uses a stratified similarity prior to balance coherence and diversity. The same engine can expand shallow inputs, such as PersonaHub entries, into deep profiles. Across three complementary evaluations, DeepPersona improves judge-extracted attribute count by 32{\%} and judge-rated uniqueness by 44{\%} over the strongest prior persona resource; improves LLM personalization quality by up to 11.6{\%} across ten dimensions, with human evaluators preferring DeepPersona responses in 81.2{\%} to 87.0{\%} of pairwise comparisons; and reduces the distributional gap to real World Values Survey responses by 31.7{\%}, while recovering Big Five personality distributions 17{\%} closer to ground truth than the strongest LLM baseline."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-deeppersona">
<titleInfo>
<title>DeepPersona: Scaling Persona Depth via Taxonomy-Guided Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zhen</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yufan</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhongyan</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lyumanshan</namePart>
<namePart type="family">Ye</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Adam</namePart>
<namePart type="family">Wood</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Saab</namePart>
<namePart type="family">Mansour</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Man</namePart>
<namePart type="family">Yao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Luoshang</namePart>
<namePart type="family">Pan</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>Synthetic personas support LLM personalization, social simulation, and alignment evaluation, yet existing resources typically represent each individual with only a small number of structured attributes. We introduce DeepPersona, a generative engine that scales persona depth to roughly 200 instantiated taxonomy nodes while preserving narrative coherence. DeepPersona combines (i) a hierarchical taxonomy of 8,496 human-descriptive attributes mined from 62k personalizable QA pairs in real user-chatbot dialogues and (ii) a progressive sampler that conditions each new attribute on the evolving persona and uses a stratified similarity prior to balance coherence and diversity. The same engine can expand shallow inputs, such as PersonaHub entries, into deep profiles. Across three complementary evaluations, DeepPersona improves judge-extracted attribute count by 32% and judge-rated uniqueness by 44% over the strongest prior persona resource; improves LLM personalization quality by up to 11.6% across ten dimensions, with human evaluators preferring DeepPersona responses in 81.2% to 87.0% of pairwise comparisons; and reduces the distributional gap to real World Values Survey responses by 31.7%, while recovering Big Five personality distributions 17% closer to ground truth than the strongest LLM baseline.</abstract>
<identifier type="citekey">wang-etal-2026-deeppersona</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2229/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>47721</start>
<end>47742</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T DeepPersona: Scaling Persona Depth via Taxonomy-Guided Generation
%A Wang, Zhen
%A Zhou, Yufan
%A Luo, Zhongyan
%A Ye, Lyumanshan
%A Wood, Adam
%A Mansour, Saab
%A Yao, Man
%A Pan, Luoshang
%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 wang-etal-2026-deeppersona
%X Synthetic personas support LLM personalization, social simulation, and alignment evaluation, yet existing resources typically represent each individual with only a small number of structured attributes. We introduce DeepPersona, a generative engine that scales persona depth to roughly 200 instantiated taxonomy nodes while preserving narrative coherence. DeepPersona combines (i) a hierarchical taxonomy of 8,496 human-descriptive attributes mined from 62k personalizable QA pairs in real user-chatbot dialogues and (ii) a progressive sampler that conditions each new attribute on the evolving persona and uses a stratified similarity prior to balance coherence and diversity. The same engine can expand shallow inputs, such as PersonaHub entries, into deep profiles. Across three complementary evaluations, DeepPersona improves judge-extracted attribute count by 32% and judge-rated uniqueness by 44% over the strongest prior persona resource; improves LLM personalization quality by up to 11.6% across ten dimensions, with human evaluators preferring DeepPersona responses in 81.2% to 87.0% of pairwise comparisons; and reduces the distributional gap to real World Values Survey responses by 31.7%, while recovering Big Five personality distributions 17% closer to ground truth than the strongest LLM baseline.
%U https://aclanthology.org/2026.findings-emnlp.2229/
%P 47721-47742
Markdown (Informal)
[DeepPersona: Scaling Persona Depth via Taxonomy-Guided Generation](https://aclanthology.org/2026.findings-emnlp.2229/) (Wang et al., Findings 2026)
ACL
- Zhen Wang, Yufan Zhou, Zhongyan Luo, Lyumanshan Ye, Adam Wood, Saab Mansour, Man Yao, and Luoshang Pan. 2026. DeepPersona: Scaling Persona Depth via Taxonomy-Guided Generation. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 47721–47742, Budapest, Hungary. Association for Computational Linguistics.