@inproceedings{wang-etal-2025-inreact,
title = "{INREACT}: An Inspire-Then-Reinforce Training Framework For Multimodal {GUI} Agent",
author = "Wang, Yuanlei and
Zhang, Liuzhou and
Luo, Haohao and
Shen, Ying",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-emnlp.486/",
doi = "10.18653/v1/2025.findings-emnlp.486",
pages = "9148--9160",
ISBN = "979-8-89176-335-7",
abstract = "Graphical User Interface (GUI) interaction, which aims to develop an intelligent GUI agent that executes user instructions to perform tasks such as installing applications by controlling digital devices, has gained significant attention due to its practical value. Although current advanced multimodal large language models (LLMs) provide GUI agents with robust perception and reasoning capabilities, they often struggle with the precise localization of small elements. To tackle this problem, we propose InReAct, a multimodal GUI agent framework that unifies observing, thinking, and acting for precise and interpretable decision-making. It is trained via a two-stage process: curriculum learning to progressively build perception, grounding, and reasoning abilities, followed by reinforcement learning to refine pixel-level grounding with an outcome-based reward. We introduce a rule-based reward function that jointly optimizes action-type selection and pixel-level localization accuracy. Experimental results on multiple datasets demonstrate the superiority of InReAct in both grounding and navigation tasks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2025-inreact">
<titleInfo>
<title>INREACT: An Inspire-Then-Reinforce Training Framework For Multimodal GUI Agent</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuanlei</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Liuzhou</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haohao</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ying</namePart>
<namePart type="family">Shen</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>Findings of the Association for Computational Linguistics: EMNLP 2025</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">979-8-89176-335-7</identifier>
</relatedItem>
<abstract>Graphical User Interface (GUI) interaction, which aims to develop an intelligent GUI agent that executes user instructions to perform tasks such as installing applications by controlling digital devices, has gained significant attention due to its practical value. Although current advanced multimodal large language models (LLMs) provide GUI agents with robust perception and reasoning capabilities, they often struggle with the precise localization of small elements. To tackle this problem, we propose InReAct, a multimodal GUI agent framework that unifies observing, thinking, and acting for precise and interpretable decision-making. It is trained via a two-stage process: curriculum learning to progressively build perception, grounding, and reasoning abilities, followed by reinforcement learning to refine pixel-level grounding with an outcome-based reward. We introduce a rule-based reward function that jointly optimizes action-type selection and pixel-level localization accuracy. Experimental results on multiple datasets demonstrate the superiority of InReAct in both grounding and navigation tasks.</abstract>
<identifier type="citekey">wang-etal-2025-inreact</identifier>
<identifier type="doi">10.18653/v1/2025.findings-emnlp.486</identifier>
<location>
<url>https://aclanthology.org/2025.findings-emnlp.486/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>9148</start>
<end>9160</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T INREACT: An Inspire-Then-Reinforce Training Framework For Multimodal GUI Agent
%A Wang, Yuanlei
%A Zhang, Liuzhou
%A Luo, Haohao
%A Shen, Ying
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Findings of the Association for Computational Linguistics: EMNLP 2025
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-335-7
%F wang-etal-2025-inreact
%X Graphical User Interface (GUI) interaction, which aims to develop an intelligent GUI agent that executes user instructions to perform tasks such as installing applications by controlling digital devices, has gained significant attention due to its practical value. Although current advanced multimodal large language models (LLMs) provide GUI agents with robust perception and reasoning capabilities, they often struggle with the precise localization of small elements. To tackle this problem, we propose InReAct, a multimodal GUI agent framework that unifies observing, thinking, and acting for precise and interpretable decision-making. It is trained via a two-stage process: curriculum learning to progressively build perception, grounding, and reasoning abilities, followed by reinforcement learning to refine pixel-level grounding with an outcome-based reward. We introduce a rule-based reward function that jointly optimizes action-type selection and pixel-level localization accuracy. Experimental results on multiple datasets demonstrate the superiority of InReAct in both grounding and navigation tasks.
%R 10.18653/v1/2025.findings-emnlp.486
%U https://aclanthology.org/2025.findings-emnlp.486/
%U https://doi.org/10.18653/v1/2025.findings-emnlp.486
%P 9148-9160
Markdown (Informal)
[INREACT: An Inspire-Then-Reinforce Training Framework For Multimodal GUI Agent](https://aclanthology.org/2025.findings-emnlp.486/) (Wang et al., Findings 2025)
ACL