[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120166-en":3,"doc-seo-120166-105":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":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":27,"seo_description":14,"update_tm":28,"read_time":29},120166,1099514068035,"Ezra","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","GPU-Accelerated Verification of Machine Learning Models for Power Systems","Computational tools for rigorously verifying large-scale machine learning models have advanced rapidly, especially via GPU-accelerated branch-and-bound solvers. Such tools are critical for deploying ML in safety-critical power systems, yet integrating them into power-system verification workflows is not turnkey. This work reformulates key power-system verification tasks into standard solver formats, enables simultaneous worst-case checking via an exact ReLU-layer transformation, and dualizes power-flow constraints for mathematical compatibility.","Proceedings of the 57th Hawaii International Conference on System Sciences | 2024  \nGPU-Accelerated Verification of Machine Learning Models  \nfor Power Systems  \nSamuel Chevalier∗  \nUniversity of Vermont [schevali@uvm.edu](schevali@uvm.edu)  \nIlgiz Murzakhanov Technical University of Denmark  [ilgmu@dtu.dk](ilgmu@dtu.dk)  \nSpyros Chatzivasileiadis†  \nTechnical University of Denmark [spchatz@dtu.dk](spchatz@dtu.dk)  \nAbstract  \nComputational tools for rigorously verifying the performance of large-scale machine learning (ML) models have progressed significantly in recent years. The most successful solvers employ highly specialized, GPU-accelerated branch and bound routines. Such tools are crucial for the successful deployment of machine learning applications in safety-critical systems, such as power systems. Despite their successes, however, barriers prevent out-of-the-box application of these routines to power system problems. This paper addresses this issue in three key ways. First, we reformulate several key power system verification problems into the canonical format utilized by modern verification solvers. Second, we enable the simultaneous verification of multiple verification problems (e.g., checking for the violation of all constraints simultaneously, and not by solving individual verification problems) . To achieve this, we introduce an exact transformation that converts a set of potential violations into a series of ReLU-based neural network layers. This allows verifiers to interpret these layers directly, and determine the “worst-case”violation in a single shot. Third, power system ML models often must be verified to satisfy power flow constraints. We propose a dualization procedure which encodes linear equality and inequality constraints (such as power balance constraints and line flow constraints) directly into the verification problem in a manner which is mathematically consistent with the specialized verification tools. To demonstrate these innovations, we verify problems associated with data-driven security  \n∗ SCC was supported by the HORIZON-MSCA-2021 Postdoc Fellowship Program, Project \\#101066991 – TRUST-ML.  \n† SCH was supported by the ERC Starting Grant VeriPhIED, funded by the European Research Council, Grant Agreement 949899 .  \nconstrained DC-OPF solvers. We build and test our first set of innovations using the α,β-CROWN solver, and we benchmark against Gurobi 10.0. Our contributions achieve a speedup that can exceed 100x and allow higher degrees of verification flexibility.  \nKeywords: Branch and bound, data driven modeling, DC-OPF, neural network verification, machine learning.  \n1. Introduction  \nThe ubiquity of machine learning (ML) applications has led to a surge of interest in topics related to adversarial robustness [1], performance verification [2], and safety guarantees [3, 4] . Collectively, these tools allow ML users to have a higher degree of trust in the underlying black-box models developed by learning algorithms. In order to spur the development and implementation of such tools, the EU Commission recently proposed the creation of regulations which will help assess ML risk, engender certification requirements, and enforce industry standards [5] . The National Institute of Standard and Technologies (NIST) in the US is looking to develop similar technical standards to help with ML regulation [6, 7] .  \nIn this paper, we focus on the problem of performance verification, i.e., formally verifying that a Neural Network (NN) model obeys certain input-output mapping properties. In particular, we focus on performance verification within the context of network-constrained, safety critical applications (i.e., electrical power system operation) . Performance verification tools have recently been developed for data-driven models used in a number of power system applications. Within the reinforcement learning (RL) context, authors in [8] developed analytical feasibility ellipsoids where RL solutions were","cbCaikFgcXCw3ed3","https://ap.wps.com/l/cbCaikFgcXCw3ed3","pdf",540350,1,10,"English","en",105,"# Abstract\n# Introduction\n## Performance verification and safety motivation\n## Related work on NN verification for power systems\n# Method overview\n## Reformulating verification into canonical solver format\n## Simultaneous verification via ReLU-layer transformation\n## Dualization of power-flow constraints\n# Experiments and results\n## Verification on data-driven security constrained DC-OPF\n## Solver comparison and benchmarking","[{\"question\":\"What problem does the paper address for power systems?\",\"answer\":\"It targets the difficulty of applying GPU-accelerated, branch-and-bound neural network verification routines directly to verification tasks arising in power-system problems.\"},{\"question\":\"How does the paper enable simultaneous verification of multiple constraints?\",\"answer\":\"It introduces an exact transformation that converts potential violations into a sequence of ReLU-based neural network layers, allowing verifiers to compute the worst-case violation in one shot.\"},{\"question\":\"How are power-flow constraints incorporated into the verification problem?\",\"answer\":\"The paper proposes a dualization procedure that encodes linear equality and inequality constraints, including power balance and line flow constraints, into the verification problem in a way consistent with the specialized verification tools.\"}]","GPU-Accelerated Verification of Machine Learning Models for Power Systems | PDF",1785728521,25,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"gpu-accelerated-verification-of-machine-learning-models-for-power-systems","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/gpu-accelerated-verification-of-machine-learning-models-for-power-systems/120166/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address for power systems?","Question",{"text":75,"@type":76},"It targets the difficulty of applying GPU-accelerated, branch-and-bound neural network verification routines directly to verification tasks arising in power-system problems.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the paper enable simultaneous verification of multiple constraints?",{"text":80,"@type":76},"It introduces an exact transformation that converts potential violations into a sequence of ReLU-based neural network layers, allowing verifiers to compute the worst-case violation in one shot.",{"name":82,"@type":73,"acceptedAnswer":83},"How are power-flow constraints incorporated into the verification problem?",{"text":84,"@type":76},"The paper proposes a dualization procedure that encodes linear equality and inequality constraints, including power balance and line flow constraints, into the verification problem in a way consistent with the specialized verification tools.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"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":21,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]