arXiv：cs.AI· Sean Culatana, Shang-En Huang, Kang Li·· 5 小时前AI 评分38
MRVQ：面向维度与码率弹性向量检索的单一常驻索引
MRVQ: One Resident Index for Dimension- and Rate-Elastic Vector Search
AI 导读
研究者提出 Matryoshka Residual Vector Quantization（MRVQ），一种针对冻结嵌入的后处理残差量化器，其最高码率编码可通过丢弃残差阶段降码率、丢弃嵌入坐标降维度，用单一常驻索引覆盖所有评估的（维度、码率）组合。
正文
Abstract:Dense-retrieval services must switch among embedding-prefix dimensions and index bit rates as latency, quality, and memory budgets change. Tuning a quantizer separately for each rate gives the best quality, but the retrieval tier then holds several code streams and quantizer states at once. We introduce Matryoshka Residual Vector Quantization (MRVQ), a post-hoc residual quantizer for frozen embeddings. Its maximum-rate code can be truncated two ways: dropping residual stages lowers the rate, and dropping embedding coordinates lowers the dimension. One resident artifact therefore serves every (dimension, rate) pair we evaluate. Across FiQA and NFCorpus, four embedding families, and {4, 8, 16}-byte codes, MRVQ is the lowest-RAM design we evaluate. It uses 17.8-22.0x less memory than three separately trained QINCo2 indices, and 1.89-2.02x less than a lean shared-model steelman. The saving is not free: per-rate QINCo2 is 0.026-0.107 nDCG@10 better on FiQA. But MRVQ beats PQ, OPQ, and AdANNS-OPQ at matched code size. We also evaluate a low-build-cost PCA-scalar design that attains quality comparable to RaBitQ and its extension while fitting 420x faster at the median. Finally, we report two negative results: QINCo2 collapses when trained at high rates, and a ranking-bound hypothesis misses its pre-specified acceptance criteria. MRVQ is therefore a low-memory operating point for elastic retrieval, not a universal quality winner.
| Subjects: | Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) |
| Cite as: | arXiv:2610.03651 [cs.AI] |
| (or arXiv:2610.03651v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03651 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Sean Culatana [view email]
[v1]
Fri, 2 Oct 2026 17:33:13 UTC (186 KB)
来源：arXiv：cs.AI · arxiv.org