[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121971-en":3,"doc-seo-121971-105":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},121971,4398048950312,"Violet","https://ap-avatar.wpscdn.com/avatar/400002538284de19e3c?_k=1778320343897328908",8,"Research & Report","Revisiting the Mark Conditional Independence Assumption in Neural Marked Temporal Point Processes","Learning marked temporal point process models requires jointly modeling event arrival times and their associated marks (labels/classes). Deep learning advances have improved event sequence modeling with more expressive neural temporal point process architectures, yet many assume marks are conditionally independent of event times given the process history. This work relaxes that assumption by parameterizing the mark distribution as a function of the current event time. Experiments on multiple real-world event sequence datasets show better future mark prediction than baselines, while keeping event time prediction performance essentially unchanged.","Revisiting the Mark Conditional Independence Assumption in Neural Marked Temporal Point  \nProcesses  \nTanguy Bosser and Souhaib Ben Taieb  \nUniversity of Mons-Department of Computer Science Avenue Victor Maistriau, 15, Mons-Belgium  \nAbstract. Learning marked temporal point process (TPP) models involves modeling both the event arrival times as well as their associated labels, referred to as marks. The recent introduction of deep learning techniques to the field led to better modeling of event sequences thanks to more flexible neural TPP models. However, some of these models make the assumption that event marks are independent of event times given the history of the process, which may not be valid in many applications. We relax this assumption and explicitly parametrize the mark distribution asa function of the current event time. We show that our approach achieves improved performance in predicting future marks compared to baselineson multiple real-world event sequence datasets, without affecting the performance on event time prediction.  \n1 Introduction  \nA broad range of systems are often characterized by sequences of discrete events taking place at irregular time intervals. Common examples may include users activity on a social media platform, e-commerce transactions, or earthquakes manifestations. Given past realizations of a system of interest, one may be interested in capturing the correlations among past event occurrences to enable prediction of future ones. In practice, these events are often associated to additional information, such as discrete classes, or marks, that we may wish to infer along the corresponding timestamp. Temporal Point Processes (TPP) [1] provide a powerful mathematical framework for modeling these streams of asynchronous and cross-correlated event data. However, classical parametrizations of TPP models, such as the Hawkes process [2], have often been criticized for their lack of flexibility in modeling complex event dynamics [3] . To increase the models’capacity, deep learning methods have been introduced to the field of TPP, including RNN [4] and self-attention mechanisms [5, 6] . Among these neural TPP architectures, LogNormMix [7] has proven itself to be a strong baseline in fitting the distribution of future event arrival times, often outperforming more recent architectures [8] . However, by assuming that the marks are conditionally independent of time given a process history, LogNormMix can hinder performance in capturing the dynamic of mark occurrences if this assumption is not valid. In this work, we provide a simple yet useful modification in the parametrization of LogNormMix to account for the dependence of future marks on time. Through  \nexperiments on 6 real-world datasets, we show that our approach often outperforms LogNormMix in predicting future marks, while keeping similar fitting capabilities when estimating the distribution of future arrival times.  \n2 Background and notations  \nMarked temporal point processes (MTPP) are stochastic processes whose realizations consist in sequences of n discrete events S = {ei = (ti , ki)} observed within a fixed window [0, T] . For each event ei , ti corresponds to the event arrival time with 0 ≤ t 1 \u003C ... \u003C tn ≤ T, while ki ∈ K = {1,..., K} is the associated mark, or class to which the event belongs. Note that S can be equivalently represented as {ei = (τi , ki)}, where τi = ti − ti−1 is the event inter-arrival time. We will use both representations interchangeably throughout the paper. In an MTPP, the occurrence of future arrival times and marks can be fully characterized through the conditional joint distribution f (t, k|Ht ), where Ht = { (ti , ki ∈ S)|ti \u003C t} is the process history up to time t. For clarity, we will employ the notation ’∗’ of [1] to indicate dependence on Ht , i.e. f (t, k|Ht ) = f ∗ (t, k) . Provided a parametric form of f ∗ (t, k;θ), the most common approach to learning the set of parameters θ is achieved by negative loglik","cbCaibZD0r3aG2ox","https://ap.wps.com/l/cbCaibZD0r3aG2ox","pdf",1662977,1,6,"English","en",105,"# Abstract\n# Introduction\n# Background and notations\n# Our TPP model","[{\"question\":\"What assumption about marks does the paper revisit in neural marked temporal point processes?\",\"answer\":\"It revisits the assumption that event marks are conditionally independent of event times given the history of the process.\"},{\"question\":\"How does the proposed method change the mark modeling approach?\",\"answer\":\"It relaxes the independence assumption by explicitly parameterizing the mark distribution as a function of the current event time.\"},{\"question\":\"What do experiments show about prediction performance?\",\"answer\":\"The approach improves future mark prediction compared with baselines on multiple real-world event sequence datasets, without harming event time prediction performance.\"}]","Revisiting the Mark Conditional Independence Assumption in Neural Marked Temporal Point Processes | PDF",1785808089,15,{"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},"revisiting-the-mark-conditional-independence-assumption-in-neural-marked-temporal-point-processes","",{"@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/revisiting-the-mark-conditional-independence-assumption-in-neural-marked-temporal-point-processes/121971/",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","2026-08-05","2026-08-04",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 assumption about marks does the paper revisit in neural marked temporal point processes?","Question",{"text":76,"@type":77},"It revisits the assumption that event marks are conditionally independent of event times given the history of the process.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does the proposed method change the mark modeling approach?",{"text":81,"@type":77},"It relaxes the independence assumption by explicitly parameterizing the mark distribution as a function of the current event time.",{"name":83,"@type":74,"acceptedAnswer":84},"What do experiments show about prediction performance?",{"text":85,"@type":77},"The approach improves future mark prediction compared with baselines on multiple real-world event sequence datasets, without harming event time prediction performance.","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"]