[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-327484-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-327484-en":131},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","qfedagent-quantum-enhanced-personalized-federated-learning-for-multi-agent-activity-recognition","QFedAgent: Quantum-Enhanced Personalized Federated Learning for Multi-Agent Activity Recognition","","Federated learning enables collaborative model training across distributed devices without sharing raw data, making it suitable for privacy-sensitive robotic sensing. For multi-agent activity recognition, heterogeneous and non-IID multimodal streams reduce the effectiveness of conventional FL, while classical fusion increases parameters and communication cost. QFedAgent presents a hybrid quantum-classical personalized FL framework using a variational quantum circuit fusion module to model accelerometer–gyroscope interactions with far fewer trainable parameters. Experiments on OPPORTUNITY under subject-based non-IID partitions reach 97.7% mean test accuracy.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/qfedagent-quantum-enhanced-personalized-federated-learning-for-multi-agent-activity-recognition/327484/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/qfedagent-quantum-enhanced-personalized-federated-learning-for-multi-agent-activity-recognition/327484.png","ImageObject",300,407,{"name":42,"@type":43},"Hazel","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-11-04","2026-09-21",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",14,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What problem does QFedAgent address in multi-agent federated learning?","Question",{"text":63,"@type":64},"It targets performance degradation caused by heterogeneous non-IID multimodal sensor streams and the high parameter/communication overhead of classical fusion methods.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"How does QFedAgent perform cross-modal fusion?",{"text":68,"@type":64},"It uses a variational quantum circuit fusion module that encodes accelerometer and gyroscope inputs into quantum states and leverages entanglement to capture interactions.",{"name":70,"@type":61,"acceptedAnswer":71},"What experimental results does the paper report on activity recognition?",{"text":72,"@type":64},"On the OPPORTUNITY dataset with subject-based non-IID partitions, it reports 97.7% mean test accuracy while achieving substantially lower fusion complexity.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},327484,1790375346,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,115,120,123,127],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":112,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":113,"slug":114},8,30,"research-report",{"id":116,"doc_module":4,"doc_module_name":25,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":25,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":25,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":128,"doc_module":4,"doc_module_name":25,"category_name":129,"show_sort_weight":97,"slug":130},19,"General","general",{"code":4,"msg":82,"data":132},{"doc_id":79,"user_id":133,"nickname":42,"user_avatar":134,"doc_module":4,"category_id":112,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":33,"language":140,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":12,"update_tm":144,"read_time":124,"preview_page_count":4},137441390410,"https://ap-avatar.wpscdn.com/avatar/2000252f4ab5702993?_k=1776741390130283984","This paper was accepted to IEEE MWSCAS 2026  \nQFedAgent: Quantum-Enhanced Personalized Federated Learning for Multi-Agent Activity  \nRecognition  \nQuoc Bao Phan and Tuy Tan Nguyen  \nDepartment of Electrical and Computer Engineering  \nFAMU-FSU College of Engineering, Florida State University  \nTallahassee, FL 32310, USA  \n{qp25c, [tuy.nguyen](tuy.nguyen}@fsu.edu)[}](tuy.nguyen}@fsu.edu)[@fsu.edu](tuy.nguyen}@fsu.edu)  \narXiv :2607 .02426v 1 [ cs .LG] 2 Jul 2026  \nAbstract—Federated learning (FL) enables collaborative model training across distributed devices without sharing raw data, making it suitable for privacy-sensitive robotic sensing applications. However, multi-agent systems generate heterogeneous and non-independent and identically distributed (non-IID) multimodal sensor streams that degrade conventional FL algorithms, while classical fusion modules introduce substantial parameter overhead and communication cost. This paper proposes QFedAgent, a hybrid quantum-classical personalized FL framework for multi-agent activity recognition. The approach integrates a variational quantum circuit fusion module that models accelerometer– gyroscope interactions through quantum state encoding and entanglement, requiring only 72 quantum rotation parameters versus 33K in classical multi-layer perceptron-based fusion, achieving approximately 10 × total parameter reduction. Experiments on the OPPORTUNITY dataset under subject-based nonIID partitions demonstrate 97.7% mean test accuracy, confirming that parameter-efficient quantum fusion remains competitive with conventional federated baselines.  \nIndex Terms—Federated learning, variational quantum circuits, multi-agent systems, activity recognition, non-IID data.  \nI. INTRODUCTION  \nAutonomous robotic systems are increasingly deployed in industrial automation, healthcare monitoring, and intelligent infrastructure. In many applications, agents operate as distributed teams equipped with heterogeneous sensing platforms such as inertial measurement units (IMUs), cameras, and LiDAR. As they continuously interact with their environments, these agents generate large volumes of multimodal sensory data that are geographically distributed across devices. However, centralizing such data for model training is often impractical due to communication overhead, latency constraints, and privacy concerns, particularly in safety-critical environments where reliable connectivity cannot be guaranteed [1] .  \nTo address these challenges, federated learning (FL) has emerged as a promising framework for collaborative training in distributed robotic systems [2] . In FL, agents train models locally on their own observations and share model updates with a coordinating server instead of raw data, preserving data locality while enabling collective learning [3] . However, classical methods such as FedAvg generally assume similar client data distributions, whereas robotic agents often operate under different environments, motion patterns, and sensor characteristics, leading to non-independent and identically distributed (non-IID) data that degrades convergence and weakens personalization [4] . This issue becomes even more  \npronounced in multimodal perception tasks such as activity recognition, where IMU-based accelerometer and gyroscope signals exhibit complex temporal and statistical dependencies. Although fusion methods such as attention, gating, and transformers can model these interactions effectively, they usually introduce substantial parameter overhead, increasing communication cost and further complicating aggregation across heterogeneous clients [5] .  \nTo mitigate these limitations, compact yet expressive fusion mechanisms are desirable for multimodal FL systems. Variational quantum circuits (VQCs) have recently emerged as differentiable models that can be optimized using gradientbased methods [6] . By encoding classical inputs into quantum states and exploiting entanglement operations, VQCscan represent co","cbCaimr7mTs5fN2R","https://ap.wps.com/l/cbCaimr7mTs5fN2R","pdf",757380,"English","# Abstract\n# I. INTRODUCTION\n# II. SYSTEM MODEL AND METHOD\n## A. Problem Formulation","[{\"question\":\"What problem does QFedAgent address in multi-agent federated learning?\",\"answer\":\"It targets performance degradation caused by heterogeneous non-IID multimodal sensor streams and the high parameter/communication overhead of classical fusion methods.\"},{\"question\":\"How does QFedAgent perform cross-modal fusion?\",\"answer\":\"It uses a variational quantum circuit fusion module that encodes accelerometer and gyroscope inputs into quantum states and leverages entanglement to capture interactions.\"},{\"question\":\"What experimental results does the paper report on activity recognition?\",\"answer\":\"On the OPPORTUNITY dataset with subject-based non-IID partitions, it reports 97.7% mean test accuracy while achieving substantially lower fusion complexity.\"}]","QFedAgent: Quantum-Enhanced Personalized Federated Learning for Multi-Agent Activity Recognition | PDF",1789977305]