[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83422-en":3,"doc-seo-83422-105":30,"detail-sidebar-cat-0-en-105":91},{"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":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},83422,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",8,"Research & Report","Learning Adaptive Solvers for Distributed Factor Graph Optimization on Matrix Lie Groups","Modern robotic perception relies on large-scale geometric optimization spread across multiple robots or sessions, yet existing distributed solvers are brittle due to hand-tuned parameters and are largely limited to rigid-body SE(3) pose graphs. DeepCORD introduces a learning-augmented framework for distributed factor graph optimization on general matrix Lie groups. It unfolds a parallel accelerated Riemannian optimizer into differentiable iterations and learns a self-supervised adaptive feedback policy driven by optimization phase and communication status. Experiments on real-world SE(3) pose graphs and SL(4) projective submap alignment show lower objectives than baselines across benchmarks in realistic regimes.","arXiv :2607 .08735v 1 [ cs .RO] 9 Jul 2026  \nLearning Adaptive Solvers for Distributed Factor Graph Optimization on Matrix Lie Groups  \nJaeho Shin 1 , Maani Ghaffari 1 , and Yulun Tian 1  \n1 University of Michigan  \nAbstract  \nModern robotic perception increasingly involves large-scale geometric optimization problems distributed across multiple robots or sessions. However, existing distributed solvers often depend on brittle hand tuning and primarily target rigid body pose graphs. To address this, we present DeepCORD, a learning-augmented framework for distributed factor graph optimization on general matrix Lie groups. By unfolding a parallel and accelerated Riemannian optimizer into differentiable iterations, DeepCORD learns a self-supervised feedback policy that dynamically adapts solver parameters according to the optimization phase and communication status. The resulting method enables adaptive distributed optimization over matrix Lie groups under both synchronous and asynchronous communication regimes. Extensive experiments on real-world SE(3) pose graph optimization and SL(4) projective submap alignment show that our method achieves lower objective values than existing distributed baselines on most benchmarks across realistic operating scenarios.  \n\n| Local Submap &\u003Cbr>Async. Communication\u003Cbr> |  |  | \u003Cbr>Initial Global Map |\n| --- | --- | --- | --- |\n|  |  |  | \u003Cbr>DeepCORD |\n| ⋯  | ⋯  |  |  |\n\n Robot 1  Robot 2  Robot 3  Robot 4  \nFigure 1: DeepCORD aligns feed-forward SLAM submaps from four robots through distributed SL(4) optimization, recovering a globally consistent map over a custom dataset spanning approximately 100 m.  \n1 Introduction  \nModern robotic perception backends increasingly require solving large-scale geometric optimization problems, where many noisy measurements must be fused into a globally consistent spatial estimate. Factor  \ngraphs [7, 11] provide a common abstraction for these problems and support optimization over different matrix Lie groups, including rigid motions for pose graph SLAM [41], similarity transforms for monocular reconstruction [48], and 3D homographies for recent systems based on geometric foundation models [29, 30] . While centralized nonlinear solvers remain powerful when full graph access is available, many multi-robot, multi-device, and multi-session systems naturally produce factor graphs that are partitioned across robots or devices [24, 25, 53] . In such settings, optimization must respect data locality, limited bandwidth, and potentially asynchronous communication.  \nRecent distributed solvers address this deployment setting by decomposing the backend across robots or graph partitions. Methods based on distributed optimization [16, 32, 45, 51] and probabilistic message passing [28, 34, 35] demonstrate that optimization can be achieved via local computation with periodic or asynchronous communication. Nevertheless, the practical performance of these solvers is currently highly sensitive to manually selected solver parameters, such as step sizes, penalty parameters, damping coefficients, and restart rules. Such sensitivity makes deployment brittle as tuned parameter settings fail to transfer across graph topology, noise level, or communication regime. Further, existing distributed solvers predominantly focus on rigid body SE(3) pose graphs, while recent perception systems increasingly involve optimization over broader matrix Lie groups.  \nContributions. Inspired by learning-to-optimize (L2O) frameworks [8], we propose DeepCORD, a learning-augmented framework for distributed factor graph optimization on matrix Lie groups. DeepCORD builds on CORD [45], a state-of-the-art distributed solver based on second-order Riemannian dynamics, and unfolds its iterations into a differentiable computation graph. Rather than learning black-box state updates, DeepCORD learns a local feedback policy that predicts adaptive solver parameters from local optimization context. The resulting learned","cbCaioUcQgyA0i5z","https://ap.wps.com/l/cbCaioUcQgyA0i5z","pdf",2265875,2,1,19,"English","en",105,"# Introduction\n# Related Work","[{\"question\":\"What problem does DeepCORD address in distributed robotic geometric optimization?\",\"answer\":\"DeepCORD targets the brittleness of existing distributed solvers, which depend on manually tuned parameters and often focus on rigid-body SE(3) pose graphs rather than broader matrix Lie groups.\"},{\"question\":\"How does DeepCORD adapt solver parameters during optimization?\",\"answer\":\"DeepCORD unfolds a parallel accelerated Riemannian optimizer into differentiable iterations and learns a self-supervised feedback policy that predicts adaptive parameters from local optimization context and communication status.\"},{\"question\":\"What evidence supports DeepCORD’s effectiveness?\",\"answer\":\"Experiments on real-world SE(3) pose graph optimization and SL(4) projective submap alignment show DeepCORD achieves lower objective values than distributed hand-tuned baselines across most benchmarks, including synchronous and asynchronous communication regimes.\"}]",1784187494,48,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"learning-adaptive-solvers-for-distributed-factor-graph-optimization-on-matrix-lie-groups","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,47,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":20},"https://docshare.wps.com/document/","Document",{"item":48,"name":12,"@type":43,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/learning-adaptive-solvers-for-distributed-factor-graph-optimization-on-matrix-lie-groups/83422/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-25","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does DeepCORD address in distributed robotic geometric optimization?","Question",{"text":75,"@type":76},"DeepCORD targets the brittleness of existing distributed solvers, which depend on manually tuned parameters and often focus on rigid-body SE(3) pose graphs rather than broader matrix Lie groups.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does DeepCORD adapt solver parameters during optimization?",{"text":80,"@type":76},"DeepCORD unfolds a parallel accelerated Riemannian optimizer into differentiable iterations and learns a self-supervised feedback policy that predicts adaptive parameters from local optimization context and communication status.",{"name":82,"@type":73,"acceptedAnswer":83},"What evidence supports DeepCORD’s effectiveness?",{"text":84,"@type":76},"Experiments on real-world SE(3) pose graph optimization and SL(4) projective submap alignment show DeepCORD achieves lower objective values than distributed hand-tuned baselines across most benchmarks, including synchronous and asynchronous communication regimes.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]