[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84759-en":3,"doc-seo-84759-105":29,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84759,4398048950312,"Violet","https://ap-avatar.wpscdn.com/avatar/400002538284de19e3c?_k=1778320343897328908",8,"Research & Report","Exploiting Structural Properties for Efficient Constraint-Aware HNSW Hyperparameter Tuning","Vector databases are central to modern retrieval systems, including Retrieval-Augmented Generation (RAG), where efficient Approximate Nearest Neighbor Search (ANNS) is critical. Hierarchical Navigable Small World (HNSW) is widely used for favorable recall-latency trade-offs, yet its hyperparameters jointly shape search quality, latency, build time, and index size in nonlinear ways. The work analyzes HNSW tuning as a systems problem, revealing structural regularities and proposing CHAT to prune constraint-infeasible configurations before full index construction. Experiments across datasets and engines show up to 45% higher throughput, 11% higher recall, and up to 44× faster convergence.","Exploiting Structural Properties for Efficient Constraint-Aware  \nHNSW Hyperparameter Tuning  \nGeon Choi  \nSeoul National University South Korea [cds06126@snu.ac.kr](cds06126@snu.ac.kr)  \nHoeun Lee  \nSeoul National University South Korea [hoeunlee@snu.ac.kr](hoeunlee@snu.ac.kr)  \nJaeyoung Do∗ Seoul National University South Korea [jaeyoung.do@snu.ac.kr](jaeyoung.do@snu.ac.kr)  \narXiv :2607 .04630v 1 [ cs .DB] 6 Jul 2026  \nAbstract  \nVector databases (VectorDBs) are a core component of modern retrieval systems, including Retrieval-Augmented Generation (RAG), where efficient Approximate Nearest Neighbor Search (ANNS) is critical. Among ANNS algorithms, Hierarchical Navigable Small World (HNSW) graphs are widely adopted for their strong recalllatency trade-off. However, configuring HNSW remains challenging: its hyperparameters jointly affect search quality, latency, build time, and index size in nonlinear ways, while production deployments impose strict resource and tuning-time constraints. We study HNSW hyperparameter tuning from a systems perspective and show that its configuration space exhibits strong structural regularities. Specifically, we identify monotonic, dominant unimodal, and separable relationships among search-time and construction-time parameters, which induce feasibility boundaries under performance and resource constraints. Building on this insight, we propose CHAT, a constraint-aware tuning framework for HNSW. Unlike generic black-box optimizers, CHAT exploits HNSW-specific structure to perform deterministic, sample-efficient search and prune resource-infeasible configurations before full index construction. Across multiple datasets and HNSW-based vector search engines, CHAT identifies configurations that maximize recall or throughput while satisfying constraints on accuracy, latency, build time, index size, and tuning budget. Compared to strong baselines, CHAT achieves up to 45% higher throughput or 11% higher recall, and converges up to 44× faster. These results show that principled, structure-aware tuning enables efficient and robust HNSW deployment beyond generic black-box optimization.  \nCCS Concepts  \n• Information systems → Nearest-neighbor search; Database query processing; • Computing methodologies → Machine learning; • Theory of computation → Design and analysis of algorithms.  \n∗ Corresponding Author.  \nPermission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission [and/or a fee. Request permissions from permissions@acm.org](and/or a fee. Request permissions from permissions@acm.org).  \nSIGMOD’27, Huntington Beach, CA, USA  \n© 2027 Copyright held by the owner/author(s) . Publication rights licensed to ACM. ACM ISBN 978-1-4503-XXXX-X/2018/06  \n[https://doi.org/XXXXXXX.XXXXXXX](https://doi.org/XXXXXXX.XXXXXXX)  \n nytimes  glove  sift  deep1M  youtube  \nRecall  \n(QPS Constraint Q90)  \nQPS (Recall Constraint = 0.95)  \nFigure 1: Performance differences between default and optimal HNSW hyperparameter configurations across datasets using Faiss [22]. Filled markers indicate optimal configurations, while outlined markers indicate default configurations. See Section 6. 1 for details.  \nKeywords  \nVector databases, Approximate nearest neighbor search, HNSW, Hyperparameter tuning, Constraint-aware optimization, Index construction  \nACM Reference Format:  \nGeon Choi, Hoeun Lee, and Jaeyoung Do. 2027. Exploiting Structural Properties for Efficient Constraint-Aware HNSW Hyperparameter Tuning. In Proceedings of the 2027 International Conference on Management of ","cbCaibnXx9aGuvjx","https://ap.wps.com/l/cbCaibnXx9aGuvjx","pdf",1975439,1,20,"English","en",105,"# Abstract\n# Introduction\n## RAG and Vector Databases\n## Why HNSW Tuning Is Difficult\n## HNSW Hyperparameters and Trade-offs","[{\"question\":\"Why is tuning HNSW hyperparameters important in RAG and vector database systems?\",\"answer\":\"Because retrieval quality and latency depend on how HNSW is configured, and HNSW hyperparameters jointly influence recall, throughput, build time, and index size. Default settings often fall outside dataset-specific feasible frontiers, causing constraint violations or leaving performance on the table.\"},{\"question\":\"What structural properties does the document claim to exploit for efficient tuning?\",\"answer\":\"It identifies monotonic, dominant unimodal, and separable relationships among search-time and construction-time parameters. These regularities create feasibility boundaries under performance and resource constraints.\"},{\"question\":\"How does CHAT differ from generic black-box hyperparameter optimization?\",\"answer\":\"CHAT uses HNSW-specific structure to run deterministic, sample-efficient search and to prune resource-infeasible configurations before full index construction. This allows it to maximize recall or throughput while meeting constraints on accuracy, latency, build time, index size, and tuning budget.\"}]",1784198076,50,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"exploiting-structural-properties-for-efficient-constraint-aware-hnsw-hyperparameter-tuning","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/exploiting-structural-properties-for-efficient-constraint-aware-hnsw-hyperparameter-tuning/84759/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"Why is tuning HNSW hyperparameters important in RAG and vector database systems?","Question",{"text":74,"@type":75},"Because retrieval quality and latency depend on how HNSW is configured, and HNSW hyperparameters jointly influence recall, throughput, build time, and index size. Default settings often fall outside dataset-specific feasible frontiers, causing constraint violations or leaving performance on the table.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"What structural properties does the document claim to exploit for efficient tuning?",{"text":79,"@type":75},"It identifies monotonic, dominant unimodal, and separable relationships among search-time and construction-time parameters. These regularities create feasibility boundaries under performance and resource constraints.",{"name":81,"@type":72,"acceptedAnswer":82},"How does CHAT differ from generic black-box hyperparameter optimization?",{"text":83,"@type":75},"CHAT uses HNSW-specific structure to run deterministic, sample-efficient search and to prune resource-infeasible configurations before full index construction. This allows it to maximize recall or throughput while meeting constraints on accuracy, latency, build time, index size, and tuning budget.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,113,118,121,125,128,132],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":28,"slug":112},6,"Technology","technology",{"id":114,"doc_module":4,"doc_module_name":45,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":119,"slug":120},30,"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":21,"slug":124},9,"Religion & Spirituality","religion-spirituality",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":126,"show_sort_weight":21,"slug":127},"World Cup","world-cup",{"id":129,"doc_module":4,"doc_module_name":45,"category_name":130,"show_sort_weight":129,"slug":131},10,"Lifestyle","lifestyle",{"id":133,"doc_module":4,"doc_module_name":45,"category_name":134,"show_sort_weight":105,"slug":135},19,"General","general"]