@inproceedings{liang-etal-2025-structured,
title = "Structured Preference Optimization for Vision-Language Long-Horizon Task Planning",
author = "Liang, Xiwen and
Lin, Min and
Ruan, Weiqi and
Xu, Rongtao and
Liu, Yuecheng and
Chen, Jiaqi and
Lin, Bingqian and
Zhuang, Yuzheng and
Liang, Xiaodan",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.emnlp-main.884/",
pages = "17501--17526",
ISBN = "979-8-89176-332-6",
abstract = "Existing vision-language planning methods perform well on short-horizon tasks but struggle with long-horizon reasoning in dynamic environments due to the difficulty of training models to generate high-quality reasoning processes. To address this, we propose Structured Preference Optimization (SPO), a framework that enhances reasoning and action selection for long-horizon task planning through structured evaluation and optimized training. SPO introduces: 1) Structured Preference Evaluation and Optimization, which evaluates reasoning chains across task relevance, historical consistency (as part of textual coherence), and image awareness (alignment with visual observations) to construct high-quality preference pairs; and 2) Curriculum-Guided Progressive Learning, enabling the model to adapt from simple to complex tasks, thereby improving generalization and robustness. To advance research in vision-language long-horizon task planning, we introduce ExtendaBench, a comprehensive benchmark covering 1,509 tasks across VirtualHome and Habitat 2.0, categorized into ultra-short, short, medium, and long tasks. Experimental results demonstrate that SPO significantly improves reasoning quality and final decision accuracy, outperforming prior methods on long-horizon tasks and underscoring the effectiveness of preference-driven optimization in vision-language task planning. Specifically, SPO achieves a +5.98{\%} GCR and +4.68{\%} SR improvement in VirtualHome and a +3.30{\%} GCR and +2.11{\%} SR improvement in Habitat over the best-performing baselines."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liang-etal-2025-structured">
<titleInfo>
<title>Structured Preference Optimization for Vision-Language Long-Horizon Task Planning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xiwen</namePart>
<namePart type="family">Liang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Min</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weiqi</namePart>
<namePart type="family">Ruan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rongtao</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuecheng</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiaqi</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bingqian</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuzheng</namePart>
<namePart type="family">Zhuang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaodan</namePart>
<namePart type="family">Liang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-332-6</identifier>
</relatedItem>
<abstract>Existing vision-language planning methods perform well on short-horizon tasks but struggle with long-horizon reasoning in dynamic environments due to the difficulty of training models to generate high-quality reasoning processes. To address this, we propose Structured Preference Optimization (SPO), a framework that enhances reasoning and action selection for long-horizon task planning through structured evaluation and optimized training. SPO introduces: 1) Structured Preference Evaluation and Optimization, which evaluates reasoning chains across task relevance, historical consistency (as part of textual coherence), and image awareness (alignment with visual observations) to construct high-quality preference pairs; and 2) Curriculum-Guided Progressive Learning, enabling the model to adapt from simple to complex tasks, thereby improving generalization and robustness. To advance research in vision-language long-horizon task planning, we introduce ExtendaBench, a comprehensive benchmark covering 1,509 tasks across VirtualHome and Habitat 2.0, categorized into ultra-short, short, medium, and long tasks. Experimental results demonstrate that SPO significantly improves reasoning quality and final decision accuracy, outperforming prior methods on long-horizon tasks and underscoring the effectiveness of preference-driven optimization in vision-language task planning. Specifically, SPO achieves a +5.98% GCR and +4.68% SR improvement in VirtualHome and a +3.30% GCR and +2.11% SR improvement in Habitat over the best-performing baselines.</abstract>
<identifier type="citekey">liang-etal-2025-structured</identifier>
<location>
<url>https://aclanthology.org/2025.emnlp-main.884/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>17501</start>
<end>17526</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Structured Preference Optimization for Vision-Language Long-Horizon Task Planning
%A Liang, Xiwen
%A Lin, Min
%A Ruan, Weiqi
%A Xu, Rongtao
%A Liu, Yuecheng
%A Chen, Jiaqi
%A Lin, Bingqian
%A Zhuang, Yuzheng
%A Liang, Xiaodan
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-332-6
%F liang-etal-2025-structured
%X Existing vision-language planning methods perform well on short-horizon tasks but struggle with long-horizon reasoning in dynamic environments due to the difficulty of training models to generate high-quality reasoning processes. To address this, we propose Structured Preference Optimization (SPO), a framework that enhances reasoning and action selection for long-horizon task planning through structured evaluation and optimized training. SPO introduces: 1) Structured Preference Evaluation and Optimization, which evaluates reasoning chains across task relevance, historical consistency (as part of textual coherence), and image awareness (alignment with visual observations) to construct high-quality preference pairs; and 2) Curriculum-Guided Progressive Learning, enabling the model to adapt from simple to complex tasks, thereby improving generalization and robustness. To advance research in vision-language long-horizon task planning, we introduce ExtendaBench, a comprehensive benchmark covering 1,509 tasks across VirtualHome and Habitat 2.0, categorized into ultra-short, short, medium, and long tasks. Experimental results demonstrate that SPO significantly improves reasoning quality and final decision accuracy, outperforming prior methods on long-horizon tasks and underscoring the effectiveness of preference-driven optimization in vision-language task planning. Specifically, SPO achieves a +5.98% GCR and +4.68% SR improvement in VirtualHome and a +3.30% GCR and +2.11% SR improvement in Habitat over the best-performing baselines.
%U https://aclanthology.org/2025.emnlp-main.884/
%P 17501-17526
Markdown (Informal)
[Structured Preference Optimization for Vision-Language Long-Horizon Task Planning](https://aclanthology.org/2025.emnlp-main.884/) (Liang et al., EMNLP 2025)
ACL
- Xiwen Liang, Min Lin, Weiqi Ruan, Rongtao Xu, Yuecheng Liu, Jiaqi Chen, Bingqian Lin, Yuzheng Zhuang, and Xiaodan Liang. 2025. Structured Preference Optimization for Vision-Language Long-Horizon Task Planning. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 17501–17526, Suzhou, China. Association for Computational Linguistics.