@inproceedings{cao-etal-2026-traceweaver,
title = "{T}race{W}eaver: Language-Conditioned Action Selection for Web Agents via Coupled Rollouts and Episodic Memory",
author = "Cao, Shuang and
Li, Rui and
Liu, Louise Ruihua and
Duprey, Alexandre",
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.51/",
pages = "1018--1035",
ISBN = "979-8-89176-462-0",
abstract = "Web agents must choose queries, clicks, form fills, and tool calls from language instructions while preserving constraints whose consequences may appear several steps later. We study this challenge as language-conditioned delayed-constraint tracking and present TraceWeaver, which couples linear-time latent rollouts, entity-centric episodic memory, and a learned critic for budgeted action selection. Across WebShop (500 episodes), Mind2Web (1,000-task split), WebArena (812-task split), and WorkArena (165-task enterprise split), TraceWeaver shows task-success gains of +2.7 to +4.7 over baselines under a matched test-time budget. Shared-pool controls reduce the confound between selection quality and proposal quality, while a language-phenomenon analysis concentrates the gains on tasks with dense, conflicting, or delayed constraints."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="cao-etal-2026-traceweaver">
<titleInfo>
<title>TraceWeaver: Language-Conditioned Action Selection for Web Agents via Coupled Rollouts and Episodic Memory</title>
</titleInfo>
<name type="personal">
<namePart type="given">Shuang</namePart>
<namePart type="family">Cao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rui</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Louise</namePart>
<namePart type="given">Ruihua</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alexandre</namePart>
<namePart type="family">Duprey</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>Web agents must choose queries, clicks, form fills, and tool calls from language instructions while preserving constraints whose consequences may appear several steps later. We study this challenge as language-conditioned delayed-constraint tracking and present TraceWeaver, which couples linear-time latent rollouts, entity-centric episodic memory, and a learned critic for budgeted action selection. Across WebShop (500 episodes), Mind2Web (1,000-task split), WebArena (812-task split), and WorkArena (165-task enterprise split), TraceWeaver shows task-success gains of +2.7 to +4.7 over baselines under a matched test-time budget. Shared-pool controls reduce the confound between selection quality and proposal quality, while a language-phenomenon analysis concentrates the gains on tasks with dense, conflicting, or delayed constraints.</abstract>
<identifier type="citekey">cao-etal-2026-traceweaver</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.51/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1018</start>
<end>1035</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T TraceWeaver: Language-Conditioned Action Selection for Web Agents via Coupled Rollouts and Episodic Memory
%A Cao, Shuang
%A Li, Rui
%A Liu, Louise Ruihua
%A Duprey, Alexandre
%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 cao-etal-2026-traceweaver
%X Web agents must choose queries, clicks, form fills, and tool calls from language instructions while preserving constraints whose consequences may appear several steps later. We study this challenge as language-conditioned delayed-constraint tracking and present TraceWeaver, which couples linear-time latent rollouts, entity-centric episodic memory, and a learned critic for budgeted action selection. Across WebShop (500 episodes), Mind2Web (1,000-task split), WebArena (812-task split), and WorkArena (165-task enterprise split), TraceWeaver shows task-success gains of +2.7 to +4.7 over baselines under a matched test-time budget. Shared-pool controls reduce the confound between selection quality and proposal quality, while a language-phenomenon analysis concentrates the gains on tasks with dense, conflicting, or delayed constraints.
%U https://aclanthology.org/2026.findings-emnlp.51/
%P 1018-1035
Markdown (Informal)
[TraceWeaver: Language-Conditioned Action Selection for Web Agents via Coupled Rollouts and Episodic Memory](https://aclanthology.org/2026.findings-emnlp.51/) (Cao et al., Findings 2026)
ACL