@inproceedings{wang-etal-2026-distilling,
title = "Distilling Large Embeddings via Hyperspherical Householder Quantization",
author = "Wang, Yihang and
Wu, Bin and
Su, Yueyang and
Zhang, Tianfu and
Du, Yiqi and
Yu, Lei and
Guo, Jiafeng and
Cheng, Xueqi",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.482/",
pages = "10562--10576",
ISBN = "979-8-89176-390-6",
abstract = "Large embedding models have become the backbone of modern retrieval systems, offering strong semantic representations at the cost of substantial storage and computation. While recent work explores quantizing embeddings into discrete document identifiers for generative retrieval, most existing approaches rely on Euclidean quantization, which is poorly aligned with the angular geometry induced by contrastive embedding training and often requires long identifier sequences to preserve semantic fidelity. In this work, we propose \textit{Hyperspherical Householder Quantization} (HHQ), a geometry-aware distillation method that compresses large embeddings into short discrete representations via iterative Householder transformations on the unit hypersphere. By explicitly preserving cosine similarity at each step, HHQ distills semantic structure into compact identifiers that remain faithful to the original embedding space. To support reliable generation of these identifiers, we introduce constrained supervised fine-tuning and tree-aware dynamic masking to enforce structural validity during training and inference. Experiments on NQ and MS{~}MARCO show that HHQ achieves competitive or superior retrieval performance using only five tokens per document, substantially reducing decoding cost while retaining strong semantic retrieval accuracy."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-distilling">
<titleInfo>
<title>Distilling Large Embeddings via Hyperspherical Householder Quantization</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yihang</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bin</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yueyang</namePart>
<namePart type="family">Su</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianfu</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yiqi</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lei</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiafeng</namePart>
<namePart type="family">Guo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xueqi</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-390-6</identifier>
</relatedItem>
<abstract>Large embedding models have become the backbone of modern retrieval systems, offering strong semantic representations at the cost of substantial storage and computation. While recent work explores quantizing embeddings into discrete document identifiers for generative retrieval, most existing approaches rely on Euclidean quantization, which is poorly aligned with the angular geometry induced by contrastive embedding training and often requires long identifier sequences to preserve semantic fidelity. In this work, we propose Hyperspherical Householder Quantization (HHQ), a geometry-aware distillation method that compresses large embeddings into short discrete representations via iterative Householder transformations on the unit hypersphere. By explicitly preserving cosine similarity at each step, HHQ distills semantic structure into compact identifiers that remain faithful to the original embedding space. To support reliable generation of these identifiers, we introduce constrained supervised fine-tuning and tree-aware dynamic masking to enforce structural validity during training and inference. Experiments on NQ and MS MARCO show that HHQ achieves competitive or superior retrieval performance using only five tokens per document, substantially reducing decoding cost while retaining strong semantic retrieval accuracy.</abstract>
<identifier type="citekey">wang-etal-2026-distilling</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.482/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>10562</start>
<end>10576</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Distilling Large Embeddings via Hyperspherical Householder Quantization
%A Wang, Yihang
%A Wu, Bin
%A Su, Yueyang
%A Zhang, Tianfu
%A Du, Yiqi
%A Yu, Lei
%A Guo, Jiafeng
%A Cheng, Xueqi
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F wang-etal-2026-distilling
%X Large embedding models have become the backbone of modern retrieval systems, offering strong semantic representations at the cost of substantial storage and computation. While recent work explores quantizing embeddings into discrete document identifiers for generative retrieval, most existing approaches rely on Euclidean quantization, which is poorly aligned with the angular geometry induced by contrastive embedding training and often requires long identifier sequences to preserve semantic fidelity. In this work, we propose Hyperspherical Householder Quantization (HHQ), a geometry-aware distillation method that compresses large embeddings into short discrete representations via iterative Householder transformations on the unit hypersphere. By explicitly preserving cosine similarity at each step, HHQ distills semantic structure into compact identifiers that remain faithful to the original embedding space. To support reliable generation of these identifiers, we introduce constrained supervised fine-tuning and tree-aware dynamic masking to enforce structural validity during training and inference. Experiments on NQ and MS MARCO show that HHQ achieves competitive or superior retrieval performance using only five tokens per document, substantially reducing decoding cost while retaining strong semantic retrieval accuracy.
%U https://aclanthology.org/2026.acl-long.482/
%P 10562-10576
Markdown (Informal)
[Distilling Large Embeddings via Hyperspherical Householder Quantization](https://aclanthology.org/2026.acl-long.482/) (Wang et al., ACL 2026)
ACL
- Yihang Wang, Bin Wu, Yueyang Su, Tianfu Zhang, Yiqi Du, Lei Yu, Jiafeng Guo, and Xueqi Cheng. 2026. Distilling Large Embeddings via Hyperspherical Householder Quantization. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 10562–10576, San Diego, California, United States. Association for Computational Linguistics.