@inproceedings{xu-etal-2026-causal,
title = "Causal Intervention with {LMM}s: Generative Text Facilitates Fine-Grained Pedestrian Anomaly Behavior Retrieval",
author = "Xu, Weifeng and
Huang, Shaofei and
Wang, Yaxiong and
Zheng, Zhedong",
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.225/",
pages = "4723--4738",
ISBN = "979-8-89176-462-0",
abstract = "Pedestrian anomaly behavior retrieval requires disambiguating subtle anomalous actions from routine ones using natural language. Existing cross-modal alignment methods often struggle with confounding biases: visual entanglement (dominant environments overshadowing target actions) and textual biases (limited stylistic diversity). These spurious correlations lead to erroneous matches among visually similar but semantically distinct samples. To address this, we reformulate the task from a causal intervention perspective. We propose a dual generative framework leveraging Large Multimodal Models (LMMs) to break environment-action and style-text confounders. Specifically, we synthesize causally meaningful hard negatives to decouple actions from environments, and augment positive queries with stylistically diverse rephrasings to eliminate narrative dependency. Extensive experiments on the Pedestrian Anomaly Behavior (PAB) benchmark yield consistent improvements of +1.71{\%} in Recall@1 and +1.01{\%} in mAP. Furthermore, zero-shot evaluation on the out-of-distribution UCC dataset demonstrates markedly enhanced generalization, achieving +3.23{\%} and +2.46{\%} gains in Recall@1 and mAP, respectively."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xu-etal-2026-causal">
<titleInfo>
<title>Causal Intervention with LMMs: Generative Text Facilitates Fine-Grained Pedestrian Anomaly Behavior Retrieval</title>
</titleInfo>
<name type="personal">
<namePart type="given">Weifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shaofei</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yaxiong</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhedong</namePart>
<namePart type="family">Zheng</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>Pedestrian anomaly behavior retrieval requires disambiguating subtle anomalous actions from routine ones using natural language. Existing cross-modal alignment methods often struggle with confounding biases: visual entanglement (dominant environments overshadowing target actions) and textual biases (limited stylistic diversity). These spurious correlations lead to erroneous matches among visually similar but semantically distinct samples. To address this, we reformulate the task from a causal intervention perspective. We propose a dual generative framework leveraging Large Multimodal Models (LMMs) to break environment-action and style-text confounders. Specifically, we synthesize causally meaningful hard negatives to decouple actions from environments, and augment positive queries with stylistically diverse rephrasings to eliminate narrative dependency. Extensive experiments on the Pedestrian Anomaly Behavior (PAB) benchmark yield consistent improvements of +1.71% in Recall@1 and +1.01% in mAP. Furthermore, zero-shot evaluation on the out-of-distribution UCC dataset demonstrates markedly enhanced generalization, achieving +3.23% and +2.46% gains in Recall@1 and mAP, respectively.</abstract>
<identifier type="citekey">xu-etal-2026-causal</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.225/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>4723</start>
<end>4738</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Causal Intervention with LMMs: Generative Text Facilitates Fine-Grained Pedestrian Anomaly Behavior Retrieval
%A Xu, Weifeng
%A Huang, Shaofei
%A Wang, Yaxiong
%A Zheng, Zhedong
%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 xu-etal-2026-causal
%X Pedestrian anomaly behavior retrieval requires disambiguating subtle anomalous actions from routine ones using natural language. Existing cross-modal alignment methods often struggle with confounding biases: visual entanglement (dominant environments overshadowing target actions) and textual biases (limited stylistic diversity). These spurious correlations lead to erroneous matches among visually similar but semantically distinct samples. To address this, we reformulate the task from a causal intervention perspective. We propose a dual generative framework leveraging Large Multimodal Models (LMMs) to break environment-action and style-text confounders. Specifically, we synthesize causally meaningful hard negatives to decouple actions from environments, and augment positive queries with stylistically diverse rephrasings to eliminate narrative dependency. Extensive experiments on the Pedestrian Anomaly Behavior (PAB) benchmark yield consistent improvements of +1.71% in Recall@1 and +1.01% in mAP. Furthermore, zero-shot evaluation on the out-of-distribution UCC dataset demonstrates markedly enhanced generalization, achieving +3.23% and +2.46% gains in Recall@1 and mAP, respectively.
%U https://aclanthology.org/2026.findings-emnlp.225/
%P 4723-4738
Markdown (Informal)
[Causal Intervention with LMMs: Generative Text Facilitates Fine-Grained Pedestrian Anomaly Behavior Retrieval](https://aclanthology.org/2026.findings-emnlp.225/) (Xu et al., Findings 2026)
ACL