针对嵌入向量结构，提出非均匀比特分配提升压缩效率
Quantization Beyond Uniform Bit Allocation
- 将嵌入分块后按需分配比特，突破传统均匀分配
- 低比特下召回率提升最高达18%，压缩率相同时表现更优
- 适合大规模检索系统中结构化嵌入的高效压缩
量化是应对现代模型生成巨大嵌入向量的关键技术。现有量化方法多忽略嵌入的几何结构，采用全维度均匀比特分配。然而，近期模型产生的嵌入具有显著结构特征。本文研究在固定内存预算下，可变比特分配是否能提升量化质量。提出一种简单框架：将嵌入划分为连续块，并在块间非均匀分配存储空间。采用贪心策略，分别应用于产品量化（PQ）和标量量化（SQ）。在具有马特约什卡特性（MRL）的嵌入上进行实验，结果一致显示，非均匀分配在相同存储预算下优于均匀基线。低比特场景下改进最显著，此时均匀分配对MRL嵌入尤为低效。在相同压缩率下，变量分配使PQ的召回率最高提升8%，SQ提升最高达18%。结果表明，面向结构的压缩与索引技术是未来重要方向。
原文摘要 · Abstract (English)
Quantization is a fundamental technique to handle the growing sizes of embeddings generated by modern models. Existing quantization schemes are largely embedding agnostic and allocate bits uniformly across dimensions. However, recent models produce embeddings with significant geometric structure. In this work, we investigate whether a variable bit allocation scheme can improve quantization quality under a fixed memory budget. We propose a simple variable bit allocation framework that partitions an embedding into contiguous buckets and allocates storage non-uniformly across them. Using a greedy allocation strategy, we instantiate this framework for both Product Quantization (PQ) and Scalar Quantization (SQ). We perform a series of experiments on embeddings known to have the Matryoshka property (MRL), and consistently observe that non-uniform allocations outperform uniform baselines at identical storage budgets. The largest improvements occur in the low-bit regime, where uniform allocation is particularly inefficient for MRL embeddings. At the same compression rates, variable allocation improves recall by up to 8\% for PQ and up to 18\% for SQ. Our results suggest a new direction for structure-aware compression and indexing techniques for large-scale retrieval systems.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。