arXiv cs.CL· Riku Kisako, Hayato Tsukagoshi, Ryohei Sasano·· 12 小时前AI 评分31
降维与量化组合对文本嵌入压缩的影响分析：0.1% 存储是否足够
When Is 0.1% Enough? Analyzing the Combined Effects of Dimensionality Reduction and Quantization on Text Embedding Compression
AI 导读
一项基于 MTEB 四个任务族、五个预训练嵌入模型的系统分析显示，降维与量化组合压缩能在远小于单一方法的存储占用下保持接近原始性能。在分类任务上，多个模型仅用不到原始嵌入 0.1% 的存储即保留至少 99% 的分类分数。维度与位宽的最佳平衡因任务而异，检索任务对压缩更敏感。
正文 · 原文
Abstract:Recent high-performing text embedding models often produce high-dimensional real-valued vectors, resulting in substantial storage and computational costs. To address this issue, embedding compression methods based on dimensionality reduction or quantization have been proposed; however, the effects of their combination have not been sufficiently investigated. In this paper, we systematically investigate how dimensionality reduction and quantization jointly affect embedding compression across four task families in the Massive Text Embedding Benchmark (MTEB), using five pretrained embedding models. Our results show that combining dimensionality reduction and quantization can preserve near-original performance with substantially smaller storage footprints than either method alone. In particular, for classification, several models retain at least 99% of their original classification score using less than 0.1% of the storage required by the original embeddings. The preferred balance between dimensionality and bit-width varies across tasks, with retrieval generally being more sensitive to compression.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2606.01074 [cs.CL] |
| (or arXiv:2606.01074v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2606.01074 arXiv-issued DOI via DataCite |
Submission history
From: Riku Kisako [view email]
[v1]
Sun, 31 May 2026 07:37:34 UTC (431 KB)
[v2]
Tue, 20 Oct 2026 08:06:45 UTC (729 KB)
来源：arXiv cs.CL · arxiv.org