mlsanityData to 13 Oct 2026

MEMONDEMAND: A Memory Management System for Large-Scale Enterprise Data

arXiv 2608.22141 ↗First posted 22 Aug 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Enterprise repositories are large, heteroge- neous, and continuously updated, making re- trieval difficult when efficient access, source- faithful evidence, and cross-query adaptation must be supported together. Enterprise mem- ory extends retrieval beyond the model con- text, but existing systems do not jointly address collection-specific hierarchy construction, low- cost routing, detailed evidence loading, and workload-aware memory updates at this scale. We introduce MEMONDEMAND, short for On- Demand Memory, a memory management sys- tem with three coordinated mechanisms: a dy- namic multi-level hierarchy that determines the abstraction structure and depth for each col- lection, dual memory at every hierarchy level that separates distilled routing from detailed evidence, and on-demand memory promotion that updates node priority under a bounded active-state budget. On EnterpriseRAG-Bench, MEMONDEMAND outperforms the strongest published LB#1 result at every evaluated scale from 10M tokens through the complete 618M- token collection, with gains of 12.23% at 10M and 4.66% at 618M. Results on FinanceBench, HotpotQA, and FRAMES further show strong performance across financial, multi-hop, and fact-retrieval settings. Together, these results establish MEMONDEMAND as an accurate, ef- ficient, and scalable memory solution for very large enterprise repositories across data scales, domains, and evidence requirements. Our code is available at https://github.com/ xfab-xinyuansong/MemOnDemand.git.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Builds on 20 of its 41 references that are in the map. First posted on arXiv 22 Aug 2026.

Builds on

20 of 41 references.

References · 20

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,221
  2. 2019Sentence-BERT: Sentence Embeddings using Siamese BERT-Networkscited by 9,107
  3. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,797
  4. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,716
  5. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,412
  6. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,815
  7. 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,433
  8. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,354
  9. 2024RULER: What's the Real Context Size of Your Long-Context Language Models?cited by 1,302
  10. 2025A-MEM: Agentic Memory for LLM Agentscited by 1,032
  11. 2016Efficient and robust approximate nearest neighbor search using Hierarchical Navigable Small World graphscited by 857
  12. 2024RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrievalcited by 527
8 more
  1. 2023LongLLMLingua: Accelerating and Enhancing LLMs in Long Context Scenarios via Prompt Compressioncited by 447
  2. 2024Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexitycited by 442
  3. 2023FinanceBench: A New Benchmark for Financial Question Answeringcited by 264
  4. 2024Fact, Fetch, and Reason: A Unified Evaluation of Retrieval-Augmented Generationcited by 159
  5. 2021FreshDiskANN: A Fast and Accurate Graph-Based ANN Index for Streaming Similarity Searchcited by 37
  6. 2021SPANN: Highly-efficient Billion-scale Approximate Nearest Neighbor Searchcited by 35
  7. 2023Worst-case Performance of Popular Approximate Nearest Neighbor Search Implementations: Guarantees and Limitationscited by 20
  8. 2026EnterpriseRAG-Bench: A RAG Benchmark for Company Internal Knowledgecited by 8

Built on by

Not cited yet.