@inproceedings{dong-etal-2026-disentangling,
title = "Disentangling Semantics and Popularity: Adversarial Debiasing and Asymmetric Alignment for {LLM}-Enhanced Long-Tail Sequential Recommendation",
author = "Dong, Wenzhe and
Shao, Minglai and
Peng, Qiyao",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1391/",
pages = "31758--31774",
ISBN = "979-8-89176-459-0",
abstract = "Sequential recommendation models dynamic user preferences but is constrained by the long-tail distribution of real-world interactions. While LLM-derived text embeddings provide semantic priors to alleviate data sparsity, existing adaptation methods are hindered by popularity bias during semantic projection and semantic prior erosion during sequence training. In this paper, we propose \textbf{A}dversarial \textbf{D}ebiasing and \textbf{A}symmetric \textbf{A}lignment (\textbf{ADAA}), a framework designed to alleviate these representation bottlenecks. Specifically, ADAA incorporates an adversarial debiasing mechanism to decouple interaction frequency factors from the projected embeddings, yielding debiased semantic representations. Furthermore, it employs an asymmetric anchoring strategy during the sequence training phase. This strategy restricts the optimization space of sparse items to preserve their injected semantic priors, while permitting head items to flexibly fit sequential transition patterns. Through a parameter-level injection mechanism, ADAA integrates offline-aligned semantics without introducing additional parameters to the online architecture. Extensive experiments on three real-world datasets demonstrate that ADAA improves long-tail recommendation accuracy while preserving the online inference efficiency of the original sequential backbone."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="dong-etal-2026-disentangling">
<titleInfo>
<title>Disentangling Semantics and Popularity: Adversarial Debiasing and Asymmetric Alignment for LLM-Enhanced Long-Tail Sequential Recommendation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wenzhe</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Minglai</namePart>
<namePart type="family">Shao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qiyao</namePart>
<namePart type="family">Peng</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Sequential recommendation models dynamic user preferences but is constrained by the long-tail distribution of real-world interactions. While LLM-derived text embeddings provide semantic priors to alleviate data sparsity, existing adaptation methods are hindered by popularity bias during semantic projection and semantic prior erosion during sequence training. In this paper, we propose Adversarial Debiasing and Asymmetric Alignment (ADAA), a framework designed to alleviate these representation bottlenecks. Specifically, ADAA incorporates an adversarial debiasing mechanism to decouple interaction frequency factors from the projected embeddings, yielding debiased semantic representations. Furthermore, it employs an asymmetric anchoring strategy during the sequence training phase. This strategy restricts the optimization space of sparse items to preserve their injected semantic priors, while permitting head items to flexibly fit sequential transition patterns. Through a parameter-level injection mechanism, ADAA integrates offline-aligned semantics without introducing additional parameters to the online architecture. Extensive experiments on three real-world datasets demonstrate that ADAA improves long-tail recommendation accuracy while preserving the online inference efficiency of the original sequential backbone.</abstract>
<identifier type="citekey">dong-etal-2026-disentangling</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1391/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>31758</start>
<end>31774</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Disentangling Semantics and Popularity: Adversarial Debiasing and Asymmetric Alignment for LLM-Enhanced Long-Tail Sequential Recommendation
%A Dong, Wenzhe
%A Shao, Minglai
%A Peng, Qiyao
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F dong-etal-2026-disentangling
%X Sequential recommendation models dynamic user preferences but is constrained by the long-tail distribution of real-world interactions. While LLM-derived text embeddings provide semantic priors to alleviate data sparsity, existing adaptation methods are hindered by popularity bias during semantic projection and semantic prior erosion during sequence training. In this paper, we propose Adversarial Debiasing and Asymmetric Alignment (ADAA), a framework designed to alleviate these representation bottlenecks. Specifically, ADAA incorporates an adversarial debiasing mechanism to decouple interaction frequency factors from the projected embeddings, yielding debiased semantic representations. Furthermore, it employs an asymmetric anchoring strategy during the sequence training phase. This strategy restricts the optimization space of sparse items to preserve their injected semantic priors, while permitting head items to flexibly fit sequential transition patterns. Through a parameter-level injection mechanism, ADAA integrates offline-aligned semantics without introducing additional parameters to the online architecture. Extensive experiments on three real-world datasets demonstrate that ADAA improves long-tail recommendation accuracy while preserving the online inference efficiency of the original sequential backbone.
%U https://aclanthology.org/2026.emnlp-main.1391/
%P 31758-31774
Markdown (Informal)
[Disentangling Semantics and Popularity: Adversarial Debiasing and Asymmetric Alignment for LLM-Enhanced Long-Tail Sequential Recommendation](https://aclanthology.org/2026.emnlp-main.1391/) (Dong et al., EMNLP 2026)
ACL