[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-134152-en":3,"doc-seo-134152-105":30,"detail-sidebar-cat-0-en-105":92},{"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":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},134152,962085564381,"Bintang","https://ap-avatar.wpscdn.com/davatar_6f874abed73319feea01a86fa6f0fab8",8,"Research & Report","ReLATE - Resilient Learner Selection for Multivariate Time-Series Classification Against Adversarial Attacks","Minimizing computational overhead in multivariate time-series classification, especially for deep learning, is difficult and becomes harder under adversarial attacks. ReLATE introduces a resilient learner selection framework that uses dataset similarity to find the most analogous source dataset, then applies the optimal model learned there. By maintaining multiple models across known adversarial scenarios and selecting efficiently via similarity metrics, ReLATE cuts overhead by an average of 81.2% while keeping performance within 4.2% of Oracle.","ReLATE: Resilient Learner Selection for Multivariate Time-Series Classification Against  \nAdversarial Attacks  \nCagla Ipek Kocal 1 , Onat Gungor2 , Aaron Tartz 1 , Tajana Rosing2 , Baris Aksanli 1  \n1 San Diego State University, San Diego, CA, USA  \n2University of California, San Diego, CA, USA  \n{ckocal0169, atartz0694, [baksanli](baksanli}@sdsu.edu)[}](baksanli}@sdsu.edu)[@sdsu.edu](baksanli}@sdsu.edu), {ogungor, [tajana](tajana}@ucsd.edu)[}](tajana}@ucsd.edu)[@ucsd.edu](tajana}@ucsd.edu)  \narXiv :2503 .07882v1 [ cs .LG] 10 Mar 2025  \nAbstract—Minimizing computational overhead in time-series classification, particularly in deep learning models, presents a significant challenge. This challenge is further compounded by adversarial attacks, emphasizing the need for resilient methods that ensure robust performance and efficient model selection. We introduce ReLATE, a framework that identifies robust learners based on dataset similarity, reduces computational overhead, and enhances resilience. ReLATE maintains multiple deep learning models in well-known adversarial attack scenarios, capturing model performance. ReLATE identifies the most analogous dataset to a given target using a similarity metric, then applies the optimal model from the most similar dataset. ReLATE reduces computational overhead by an average of 81.2%, enhancing adversarial resilience and streamlining robust model selection, all without sacrificing performance, within 4.2% of Oracle.  \nIndex Terms—Cyber Security, Resilient Machine Learning, Adversarial attacks, Time Series Classification  \nI. INTRODUCTION  \nVarious tasks rely on time-series data, i.e., sequences of observations collected over intervals, including anomaly detection [1], clustering [2], and classification [3] . Among these tasks, time-series classification with machine learning (ML) has crucial use cases, e.g., network intrusion detection [4], event logs classification [5], malware detection [6], epileptic activity classification using EEG signals [7], and smart agriculture using multispectral satellite imagery [8], requiring robust, resilient, secure, and accurate ML-based solutions.  \nTime-series ML applications face significant challenges due to the dynamic nature of streaming data, which is often limited or incomplete in real-time environments, making it impractical to wait for sufficient data accumulation to retrain models [9] . Moreover, training ML models on new data is both computationally expensive and time-consuming, further complicating the process [10] . In this context, deep learning (DL) models are often favored for multivariate time-series classification tasks due to their ability to automatically extract relevant features. However, these DL models could show significant variability in classification performance, as shown in Figure 1 . These results highlight the substantial impact of model choice on classification outcomes, underscoring the critical need for careful and informed model selection in the context of DL-based time-series analysis. This motivates the  \n\n|  |\n| --- |\n|  |\n|  |\n|  |\n|  |\n|  |\n\n40  \n20  \n0  \nUWaveGestureLibrary  \n120  \n100  \n80  \n60  \nERing  \nBasicMotions  \nEpilepsy  \nDataset Name  \n TCN  \nFig. 1: DL performance on multivariate time-series data  \nneed for efficient DL model selection methods that can adapt to new incoming data without requiring extensive retraining.  \nDeep learning (DL) models are also vulnerable to adversarial attacks, as they can be manipulated by small, imperceptible changes in the input data, especially when the data is limited or incomplete. These attacks introduce deliberate perturbations that obscure essential patterns, potentially leading tomisclassifications in high-stakes applications where accuracy is critical. For instance, small perturbations in medical sensor data could lead to incorrect diagnoses, potentially endangering lives, while adversarial attacks in security systems could result in unauthorized access or compromised s","cbCaigop1u1S55rN","https://ap.wps.com/l/cbCaigop1u1S55rN","pdf",397059,1,6,"English","en",105,"# I. INTRODUCTION\n## Time-series ML challenges\n## Adversarial attack vulnerability and resilience need\n# II. RELATED WORK\n## A. Multivariate Time-Series Classification (MTSC)\n## B. Similarity Based Approaches","[{\"question\":\"What problem does ReLATE address in multivariate time-series classification?\",\"answer\":\"ReLATE targets the high computational and retraining cost of selecting deep learning models for streaming multivariate time-series, particularly when adversarial attacks are present.\"},{\"question\":\"How does ReLATE choose the appropriate model for a target dataset?\",\"answer\":\"ReLATE computes dataset similarity to identify the most analogous dataset, then applies the optimal model associated with that similar dataset.\"},{\"question\":\"What performance and efficiency improvements does ReLATE achieve?\",\"answer\":\"ReLATE reduces computational overhead by about 81.2% on average while maintaining strong classification performance, staying within 4.2% of Oracle.\"}]","ReLATE - Resilient Learner Selection for Multivariate Time-Series Classification Against Adversarial Attacks | PDF",1787235225,15,{"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":87,"head_meta":89,"extra_data":91,"updated_unix":28},"relate-resilient-learner-selection-for-multivariate-time-series-classification-against-adversarial-attacks","",{"@graph":36,"@context":86},[37,54,69],{"@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/relate-resilient-learner-selection-for-multivariate-time-series-classification-against-adversarial-attacks/134152/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-23","2026-08-20",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What problem does ReLATE address in multivariate time-series classification?","Question",{"text":76,"@type":77},"ReLATE targets the high computational and retraining cost of selecting deep learning models for streaming multivariate time-series, particularly when adversarial attacks are present.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does ReLATE choose the appropriate model for a target dataset?",{"text":81,"@type":77},"ReLATE computes dataset similarity to identify the most analogous dataset, then applies the optimal model associated with that similar dataset.",{"name":83,"@type":74,"acceptedAnswer":84},"What performance and efficiency improvements does ReLATE achieve?",{"text":85,"@type":77},"ReLATE reduces computational overhead by about 81.2% on average while maintaining strong classification performance, staying within 4.2% of Oracle.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":93},[94,98,102,106,111,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},"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":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":107,"slug":138},19,"General","general"]