Spectral Tempering for Embedding Compression in Dense Passage Retrieval

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Yongkang, Eustratiadis, Panagiotis, Kanoulas, Evangelos
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914481506877440
author Li, Yongkang
Eustratiadis, Panagiotis
Kanoulas, Evangelos
author_facet Li, Yongkang
Eustratiadis, Panagiotis
Kanoulas, Evangelos
contents Dimensionality reduction is critical for deploying dense retrieval systems at scale, yet mainstream post-hoc methods face a fundamental trade-off: principal component analysis (PCA) preserves dominant variance but underutilizes representational capacity, while whitening enforces isotropy at the cost of amplifying noise in the heavy-tailed eigenspectrum of retrieval embeddings. Intermediate spectral scaling methods unify these extremes by reweighting dimensions with a power coefficient $γ$, but treat $γ$ as a fixed hyperparameter that requires task-specific tuning. We show that the optimal scaling strength $γ$ is not a global constant: it varies systematically with target dimensionality $k$ and is governed by the signal-to-noise ratio (SNR) of the retained subspace. Based on this insight, we propose Spectral Tempering (\textbf{SpecTemp}), a learning-free method that derives an adaptive $γ(k)$ directly from the corpus eigenspectrum using local SNR analysis and knee-point normalization, requiring no labeled data or validation-based search. Extensive experiments demonstrate that Spectral Tempering consistently achieves near-oracle performance relative to grid-searched $γ^*(k)$ while remaining fully learning-free and model-agnostic. Our code is publicly available at https://github.com/liyongkang123/SpecTemp.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19339
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Spectral Tempering for Embedding Compression in Dense Passage Retrieval
Li, Yongkang
Eustratiadis, Panagiotis
Kanoulas, Evangelos
Information Retrieval
Artificial Intelligence
Computation and Language
Dimensionality reduction is critical for deploying dense retrieval systems at scale, yet mainstream post-hoc methods face a fundamental trade-off: principal component analysis (PCA) preserves dominant variance but underutilizes representational capacity, while whitening enforces isotropy at the cost of amplifying noise in the heavy-tailed eigenspectrum of retrieval embeddings. Intermediate spectral scaling methods unify these extremes by reweighting dimensions with a power coefficient $γ$, but treat $γ$ as a fixed hyperparameter that requires task-specific tuning. We show that the optimal scaling strength $γ$ is not a global constant: it varies systematically with target dimensionality $k$ and is governed by the signal-to-noise ratio (SNR) of the retained subspace. Based on this insight, we propose Spectral Tempering (\textbf{SpecTemp}), a learning-free method that derives an adaptive $γ(k)$ directly from the corpus eigenspectrum using local SNR analysis and knee-point normalization, requiring no labeled data or validation-based search. Extensive experiments demonstrate that Spectral Tempering consistently achieves near-oracle performance relative to grid-searched $γ^*(k)$ while remaining fully learning-free and model-agnostic. Our code is publicly available at https://github.com/liyongkang123/SpecTemp.
title Spectral Tempering for Embedding Compression in Dense Passage Retrieval
topic Information Retrieval
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2603.19339