[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84312-en":3,"doc-seo-84312-105":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},84312,687197207919,"Theodora","https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552",6,"Technology","zkComposer: Decomposing Proof Construction to Scale zkML","Zero-knowledge machine learning (zkML) lets a server provide verifiable inference while keeping model parameters private from the client. Existing zkML systems suffer from prohibitive proof-generation costs because prover time does not scale well with additional threads: proofs are built monolithically for the entire model. zkComposer introduces modular proof construction that decomposes correct-inference proofs into independent sub-proofs linked via shared activation boundary commitments, preserving monolithic guarantees without extra linking proofs or cryptographic changes.","zkComposer: Decomposing Proof Construction to Scale zkML  \nPawan Kumar Sanjaya∗ , Christina Giannoula†§ , Valdy Oktavian∗ , Mehdi Saeedi‡, Gabor Sines‡,  \nGururaj Saileshwar∗ , Nandita Vijaykumar∗  \n∗ University of Toronto, Canada  \n†Max Planck Institute for Software Systems (MPI-SWS), Germany  \n‡Advanced Micro Devices (AMD)  \narXiv :2607 .08095v 1 [ cs .CR] 9 Jul 2026  \nAbstract—Zero-knowledge machine learning (zkML) enables a server to perform verifiable inference while keeping model parameters private from the client. However, existing zkML systems incur prohibitive proof-generation costs. We observe that proof generation exhibits limited parallelism; that is, prover time does not decrease significantly as the number of threads increases. This limitation is because existing systems rely on monolithic proof computation, constructing a single proof for the entire machine learning model.  \nWe introduce zkComposer, a modular proof-construction framework that unlocks an additional dimension of parallelism, in addition to the parallelism in existing proof kernels. zkComposer decomposes the zkML proof of correct inference into independent sub-proofs, each covering a subset of the computation for inference, e.g., each independent sub-proof can cover a subset of contiguous layers in the ML model. Adjacent sub-proofs are cryptographically linked through shared commitments to the activations from the boundary layer. zkComposer provides the same guarantees as the monolithic proof without requiring additional linking proofs or changes to the underlying cryptographic primitives.  \nWe implement zkComposer and evaluate it on three CNNsand GPT-2. We show that, on CNN workloads, zkComposer reduces prover time and response time by up to 3.25 × relative to zkCNN [1]. On GPT-2, zkComposer reduces these times by up to 4.83 × relative to zkGPT [2], when partitioning along the model layers. When partitioning across both model layers and input sequences in GPT-2, we show that zkComposer reduces prover time and response time by up to 6.84 × relative to zkGPT [2].  \n1. Introduction  \nMachine learning (ML) inference systems are increas  \ningly deployed in high-stakes domains including personalized recommendations [3], fraud detection [4], insurance § Work partially conducted while the author was at the University of Toronto.  \n© 2026 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD Arrow logo, Radeon, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other product names used in this publication are for identification purposes only and may be trademarks of their respective companies.  \nclaim processing [5], and healthcare diagnostics [6] . In such settings, it is often critical that a response is generated by a specific model. Such a model may have been verified to satisfy formal properties such as fairness [7], or certified to meet a required quality of output. These guarantees do not hold if a different model is substituted. However, it is difficult to ensure that the service providers who host and deploy these models actually use the specified model [8] . For example, a provider could substitute a cheaper model in place of the one specified by the user. Revealing the model weights would let users rerun the inference and validate the response. However, weights are trade secrets, requiring significant financial resources to train (e.g., $40M for GPT- 4 [9]) and cannot be revealed [1], [2], [10] . Thus there is a need for a mechanism that lets users verify that a response was generated by the specified model, without knowing the model weights.  \nRecent advances in cryptography, particularly zeroknowledge proofs (ZKPs) [11], [12], [13], offer a practical mechanism to enforce these guarantees. A ZKP allows a prover to convince a verifier that a given computation was executed correctly without revealing any private inputs, intermediate values, or other sensitive information beyond what is implied by the output alone [14","cbCaimS2vcOrsiNZ","https://ap.wps.com/l/cbCaimS2vcOrsiNZ","pdf",645420,1,19,"English","en",105,"# Abstract\n# Introduction\n## Verifiable inference in zkML\n## ZKPs and circuit-based correctness\n## zkML threat model and privacy goals\n## Proof generation cost challenges\n# Figure overview","[{\"question\":\"What problem does zkComposer address in existing zkML systems?\",\"answer\":\"Existing zkML systems make proof generation too expensive because proof computation is largely monolithic, so prover time does not decrease much as thread count increases.\"},{\"question\":\"How does zkComposer improve proof generation for zkML?\",\"answer\":\"It decomposes the zkML proof into independent sub-proofs covering subsets of the model computation (e.g., contiguous layers) and cryptographically links adjacent sub-proofs using shared commitments to boundary-layer activations.\"},{\"question\":\"What performance improvements does zkComposer report on evaluated models?\",\"answer\":\"On CNN workloads it reduces prover time and response time by up to 3.25× versus zkCNN, while on GPT-2 it reduces these metrics by up to 4.83× (layer partitioning) and up to 6.84× (layer and input-sequence partitioning) versus zkGPT.\"}]",1784194745,48,{"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},"zkcomposer-decomposing-proof-construction-to-scale-zkml","",{"@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/technology/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/zkcomposer-decomposing-proof-construction-to-scale-zkml/84312/",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},"What problem does zkComposer address in existing zkML systems?","Question",{"text":74,"@type":75},"Existing zkML systems make proof generation too expensive because proof computation is largely monolithic, so prover time does not decrease much as thread count increases.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does zkComposer improve proof generation for zkML?",{"text":79,"@type":75},"It decomposes the zkML proof into independent sub-proofs covering subsets of the model computation (e.g., contiguous layers) and cryptographically links adjacent sub-proofs using shared commitments to boundary-layer activations.",{"name":81,"@type":72,"acceptedAnswer":82},"What performance improvements does zkComposer report on evaluated models?",{"text":83,"@type":75},"On CNN workloads it reduces prover time and response time by up to 3.25× versus zkCNN, while on GPT-2 it reduces these metrics by up to 4.83× (layer partitioning) and up to 6.84× (layer and input-sequence partitioning) versus zkGPT.","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},[92,96,100,104,109,112,117,122,127,130,134],{"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":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":110,"slug":111},50,"technology",{"id":113,"doc_module":4,"doc_module_name":45,"category_name":114,"show_sort_weight":115,"slug":116},7,"Healthcare",40,"healthcare",{"id":118,"doc_module":4,"doc_module_name":45,"category_name":119,"show_sort_weight":120,"slug":121},8,"Research & Report",30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},"General","general"]