arXiv cs.AI· Hilaf Hasson, Aditya Chakravarty, Jayant Thomas, Krishna Gogineni·· 13 小时前AI 评分55
ScalableRAG：以极低摄入成本实现高质量 RAG 与精确语料操作
ScalableRAG: High-Quality RAG with Exact Corpus Operations at Minimal Ingestion Cost
AI 导读
ScalableRAG 提出持久化类型集合工作区，让智能体在查询时构建文档集与值集，并通过集合代数、类型化抽取和聚合完成精确的语料级操作。零摄入版本无需预处理甚至不需要向量索引，在四个常规 QA 基准上与领先基线持平（74.75 对 74.64），在 Hotels-50 上得分 71.74，比领先基线 SRAG 高 36.23 分；去掉工作区后该分数降至 38.4。
正文 · 原文
Abstract:RAG systems face a capability-cost gap: lightweight retrievers locate evidence but expose no operands for exact corpus-level operations (counts, intersections, aggregations), while knowledge ingestion systems obtain them only through expensive ingestion. ScalableRAG closes this gap with a persistent typed-set workspace: an agent builds named document and value sets at query time and composes them through set algebra, typed extraction, aggregation, and diagnostics.
Zero-Ingestion ScalableRAG needs no preprocessing, not even a vector index. It matches the leading baseline on four conventional QA benchmarks (74.75 versus 74.64) and on aggregative Hotels-50 beats every evaluated system, scoring 71.74: 36.23 above the leading baseline (SRAG), and 40.22 points above A-RAG. Removing the workspace drops Hotels-50 to 38.4. On Hotels-350, the partition SRAG itself reports on, ScalableRAG leads SRAG by 9.3 points. Limited-Ingestion ScalableRAG adds a lightweight index and is the only evaluated system above 80 judge score at 5M documents, at 13 to 16 seconds per question and under 20 minutes ingestion; at 1M it scores 84.5 to A-RAG's 80.0 with 16x lower latency and 0.5% of its ingestion time. ScalableRAG combines minimal ingestion, conventional-QA parity, exact corpus operations, and million-document viability; no evaluated alternative provides all four.
Our code is available at this https URL .
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| ACM classes: | I.2.7; H.3.3; H.3.4; I.2.4 |
| Cite as: | arXiv:2607.25135 [cs.AI] |
| (or arXiv:2607.25135v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2607.25135 arXiv-issued DOI via DataCite |
Submission history
From: Hilaf Hasson [view email]
[v1]
Mon, 27 Jul 2026 23:03:30 UTC (59 KB)
[v2]
Mon, 26 Oct 2026 23:52:30 UTC (77 KB)
来源：arXiv cs.AI · arxiv.org