@inproceedings{kumar-etal-2026-enhancing,
title = "Enhancing User Safety: Context-Aware Detection of Offensive Query-Ad Pairs in Multimodal Search Advertising",
author = "Kumar, Gaurav and
Xi, Qiangjian and
Dabral, Tanmaya Shekhar and
Ghasemi, Hooshang and
Krishnamoorthy, Abishek and
Fu, Danqing and
Min, Rui and
Antunez, Emilio and
Ding, Zhongli and
Narayana, Pradyumna",
editor = {Matusevych, Yevgen and
Eryi{\u{g}}it, G{\"u}l{\c{s}}en and
Aletras, Nikolaos},
booktitle = "Proceedings of the 19th Conference of the {E}uropean Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 5: Industry Track)",
month = mar,
year = "2026",
address = "Rabat, Morocco",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.eacl-industry.36/",
pages = "475--482",
ISBN = "979-8-89176-384-5",
abstract = "The proliferation of multi-modal online advertisements necessitates robust content moderation to ensure user safety, as offensive ad content can cause user distress and erode platform trust. This paper addresses the detection of content that becomes offensive only when a user{'}s search query is paired with a specific ad, a context-dependent challenge that simple moderation often misses. Key challenges include the nuanced, multi-modal nature of ads, severe data scarcity and class imbalance due to the rarity of offensive content, and the high cost of human labeling. To overcome these limitations, we introduce a novel, context-aware detection framework centered on a large-scale, Multi-modal Teacher-Student Knowledge Distillation architecture. A powerful Gemini encoder-only ``teacher'' model distills its knowledge into a lightweight student model suitable for low-latency deployment. We enhance robustness using a novel graph mining technique to find rare offensive examples for training. For evaluation, we developed a highly accurate Automated Evaluation Model (AEM){---}a separate, larger Gemini model utilizing Chain-of-Thought (CoT) reasoning{---}to rigorously assess performance in a live A/B test. Our results demonstrate that the proposed framework reduces the serving of offensive query-ad pairs by more than 80{\%} compared to the baseline, while maintaining the efficiency required for real-time advertising systems that operate at a scale of over $\approx100$ billion query-ad pairs per day. Disclaimer: This paper contains sentences and images that may be offensive. These examples are included solely for scientific analysis and do not reflect the views of the authors."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kumar-etal-2026-enhancing">
<titleInfo>
<title>Enhancing User Safety: Context-Aware Detection of Offensive Query-Ad Pairs in Multimodal Search Advertising</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gaurav</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qiangjian</namePart>
<namePart type="family">Xi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmaya</namePart>
<namePart type="given">Shekhar</namePart>
<namePart type="family">Dabral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hooshang</namePart>
<namePart type="family">Ghasemi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Abishek</namePart>
<namePart type="family">Krishnamoorthy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Danqing</namePart>
<namePart type="family">Fu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rui</namePart>
<namePart type="family">Min</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Emilio</namePart>
<namePart type="family">Antunez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhongli</namePart>
<namePart type="family">Ding</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pradyumna</namePart>
<namePart type="family">Narayana</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-03</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yevgen</namePart>
<namePart type="family">Matusevych</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gülşen</namePart>
<namePart type="family">Eryiğit</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikolaos</namePart>
<namePart type="family">Aletras</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Rabat, Morocco</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-384-5</identifier>
</relatedItem>
<abstract>The proliferation of multi-modal online advertisements necessitates robust content moderation to ensure user safety, as offensive ad content can cause user distress and erode platform trust. This paper addresses the detection of content that becomes offensive only when a user’s search query is paired with a specific ad, a context-dependent challenge that simple moderation often misses. Key challenges include the nuanced, multi-modal nature of ads, severe data scarcity and class imbalance due to the rarity of offensive content, and the high cost of human labeling. To overcome these limitations, we introduce a novel, context-aware detection framework centered on a large-scale, Multi-modal Teacher-Student Knowledge Distillation architecture. A powerful Gemini encoder-only “teacher” model distills its knowledge into a lightweight student model suitable for low-latency deployment. We enhance robustness using a novel graph mining technique to find rare offensive examples for training. For evaluation, we developed a highly accurate Automated Evaluation Model (AEM)—a separate, larger Gemini model utilizing Chain-of-Thought (CoT) reasoning—to rigorously assess performance in a live A/B test. Our results demonstrate that the proposed framework reduces the serving of offensive query-ad pairs by more than 80% compared to the baseline, while maintaining the efficiency required for real-time advertising systems that operate at a scale of over \approx100 billion query-ad pairs per day. Disclaimer: This paper contains sentences and images that may be offensive. These examples are included solely for scientific analysis and do not reflect the views of the authors.</abstract>
<identifier type="citekey">kumar-etal-2026-enhancing</identifier>
<location>
<url>https://aclanthology.org/2026.eacl-industry.36/</url>
</location>
<part>
<date>2026-03</date>
<extent unit="page">
<start>475</start>
<end>482</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Enhancing User Safety: Context-Aware Detection of Offensive Query-Ad Pairs in Multimodal Search Advertising
%A Kumar, Gaurav
%A Xi, Qiangjian
%A Dabral, Tanmaya Shekhar
%A Ghasemi, Hooshang
%A Krishnamoorthy, Abishek
%A Fu, Danqing
%A Min, Rui
%A Antunez, Emilio
%A Ding, Zhongli
%A Narayana, Pradyumna
%Y Matusevych, Yevgen
%Y Eryiğit, Gülşen
%Y Aletras, Nikolaos
%S Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)
%D 2026
%8 March
%I Association for Computational Linguistics
%C Rabat, Morocco
%@ 979-8-89176-384-5
%F kumar-etal-2026-enhancing
%X The proliferation of multi-modal online advertisements necessitates robust content moderation to ensure user safety, as offensive ad content can cause user distress and erode platform trust. This paper addresses the detection of content that becomes offensive only when a user’s search query is paired with a specific ad, a context-dependent challenge that simple moderation often misses. Key challenges include the nuanced, multi-modal nature of ads, severe data scarcity and class imbalance due to the rarity of offensive content, and the high cost of human labeling. To overcome these limitations, we introduce a novel, context-aware detection framework centered on a large-scale, Multi-modal Teacher-Student Knowledge Distillation architecture. A powerful Gemini encoder-only “teacher” model distills its knowledge into a lightweight student model suitable for low-latency deployment. We enhance robustness using a novel graph mining technique to find rare offensive examples for training. For evaluation, we developed a highly accurate Automated Evaluation Model (AEM)—a separate, larger Gemini model utilizing Chain-of-Thought (CoT) reasoning—to rigorously assess performance in a live A/B test. Our results demonstrate that the proposed framework reduces the serving of offensive query-ad pairs by more than 80% compared to the baseline, while maintaining the efficiency required for real-time advertising systems that operate at a scale of over \approx100 billion query-ad pairs per day. Disclaimer: This paper contains sentences and images that may be offensive. These examples are included solely for scientific analysis and do not reflect the views of the authors.
%U https://aclanthology.org/2026.eacl-industry.36/
%P 475-482
Markdown (Informal)
[Enhancing User Safety: Context-Aware Detection of Offensive Query-Ad Pairs in Multimodal Search Advertising](https://aclanthology.org/2026.eacl-industry.36/) (Kumar et al., EACL 2026)
ACL
- Gaurav Kumar, Qiangjian Xi, Tanmaya Shekhar Dabral, Hooshang Ghasemi, Abishek Krishnamoorthy, Danqing Fu, Rui Min, Emilio Antunez, Zhongli Ding, and Pradyumna Narayana. 2026. Enhancing User Safety: Context-Aware Detection of Offensive Query-Ad Pairs in Multimodal Search Advertising. In Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track), pages 475–482, Rabat, Morocco. Association for Computational Linguistics.