[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-1-en-105":3,"doc-seo-260477-105":53,"doc-detail-260477-en":126},{"code":4,"msg":5,"data":6},0,"success",[7,14,19,24,29,34,39,44,49],{"id":8,"doc_module":9,"doc_module_name":10,"category_name":11,"show_sort_weight":12,"slug":13},11,1,"Template","Presentations",90,"presentations",{"id":15,"doc_module":9,"doc_module_name":10,"category_name":16,"show_sort_weight":17,"slug":18},12,"Resumes",80,"resumes",{"id":20,"doc_module":9,"doc_module_name":10,"category_name":21,"show_sort_weight":22,"slug":23},14,"Invoices",70,"invoices",{"id":25,"doc_module":9,"doc_module_name":10,"category_name":26,"show_sort_weight":27,"slug":28},15,"Posters",60,"posters",{"id":30,"doc_module":9,"doc_module_name":10,"category_name":31,"show_sort_weight":32,"slug":33},16,"Social Media",50,"social-media",{"id":35,"doc_module":9,"doc_module_name":10,"category_name":36,"show_sort_weight":37,"slug":38},17,"Forms",40,"forms",{"id":40,"doc_module":9,"doc_module_name":10,"category_name":41,"show_sort_weight":42,"slug":43},18,"Letters",30,"letters",{"id":45,"doc_module":9,"doc_module_name":10,"category_name":46,"show_sort_weight":47,"slug":48},21,"Paper Templates",5,"papers-templates",{"id":50,"doc_module":9,"doc_module_name":10,"category_name":51,"show_sort_weight":4,"slug":52},158,"General","general-158",{"code":4,"msg":54,"data":55},"ok",{"site_id":56,"language":57,"slug":58,"title":59,"keywords":60,"description":61,"schema_data":62,"social_meta":119,"head_meta":121,"extra_data":123,"updated_unix":125},105,"en","learning-job-title-representation-from-job-description-aggregation","Learning Job Title Representation from Job Description Aggregation","","Learning job title representation is essential for building automated human resource tools, yet prior approaches often learn title representations primarily from skills extracted from job descriptions and ignore richer, diverse information inside the text. This work proposes a JD-based framework that learns job titles from their job descriptions using a Job Description Aggregator to manage long text and a bidirectional contrastive loss to model the two-way relationship between titles and descriptions, outperforming skill-based baselines across in-domain and out-of-domain settings.",{"@graph":63,"@context":118},[64,80,101],{"@type":65,"itemListElement":66},"BreadcrumbList",[67,71,74,77],{"item":68,"name":69,"@type":70,"position":9},"https://docshare.wps.com","Home","ListItem",{"item":72,"name":10,"@type":70,"position":73},"https://docshare.wps.com/template/",2,{"item":75,"name":51,"@type":70,"position":76},"https://docshare.wps.com/template/general/",3,{"item":78,"name":59,"@type":70,"position":79},"https://docshare.wps.com/template/learning-job-title-representation-from-job-description-aggregation/260477/",4,{"url":78,"name":59,"@type":81,"image":82,"author":87,"headline":59,"publisher":90,"fileFormat":93,"inLanguage":57,"description":61,"dateModified":94,"datePublished":95,"encodingFormat":93,"isAccessibleForFree":96,"interactionStatistic":97},"DigitalDocument",{"url":83,"@type":84,"width":85,"height":86},"https://docshare.wps.com/thumbnails/learning-job-title-representation-from-job-description-aggregation/260477.png","ImageObject",442,249,{"name":88,"@type":89},"Jordan Avery","Person",{"url":68,"name":91,"@type":92},"DocShare","Organization","application/pdf","2026-09-21","2026-09-13",true,{"@type":98,"interactionType":99,"userInteractionCount":76},"InteractionCounter",{"@type":100},"ViewAction",{"@type":102,"mainEntity":103},"FAQPage",[104,110,114],{"name":105,"@type":106,"acceptedAnswer":107},"Why is learning job title representation important in recruitment systems?","Question",{"text":108,"@type":109},"It enables automation of job-related tasks such as job recommendation, job trajectory prediction, and benchmarking, by helping systems understand post semantics, especially job titles.","Answer",{"name":111,"@type":106,"acceptedAnswer":112},"What is the main limitation of prior skill-based approaches?",{"text":113,"@type":109},"They depend on skill information, which can be manually listed with errors or be extracted via pipelines that require predefined vocabularies or curated datasets and must be kept up to date.",{"name":115,"@type":106,"acceptedAnswer":116},"How does the proposed method learn job title representations without skill extraction?",{"text":117,"@type":109},"It learns from job descriptions by using a Job Description Aggregator network to reweight JD segments into a unified representation, then applies bidirectional contrastive loss to align job titles with their aggregated descriptions.","https://schema.org",{"og:url":78,"og:type":120,"og:title":59,"og:site_name":91,"og:description":61},"article",{"robots":122,"canonical":78},"index,follow",{"doc_id":124,"site_id":56},260477,1789342400,{"code":4,"msg":5,"data":127},{"doc_id":124,"user_id":128,"nickname":88,"user_avatar":129,"doc_module":9,"category_id":50,"category_name":51,"doc_title":59,"doc_description":61,"doc_content":130,"file_id":131,"file_url":132,"file_type":133,"file_size":134,"view_count":76,"is_deleted":4,"is_public":9,"is_downloadable":9,"audit_status":9,"page_count":8,"language":135,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":136,"faqs":137,"seo_title":138,"seo_description":61,"update_tm":125,"read_time":79},1099523882367,"https://ap-avatar.wpscdn.com/davatar_9964176cb1d06d4a9deccf72a44ae3dc","Learning Job Title Representation from Job Description Aggregation  \nNetwork  \nNapat Laosaengpha♠ , Thanit Tativannarat♠ , Chawan Piansaddhayanon♡ Attapol Rutherford♢ , and Ekapol Chuangsuwanich♠, ♡♠Department of Computer Engineering, Faculty of Engineering, Chulalongkorn University ♡ Center of Excellence in Computational Molecular Biology,  \nFaculty of Medicine, Chulalongkorn University  \n♢Department of Linguistics, Faculty of Arts, Chulalongkorn University, Thailand  \n[napatnicky@gmail.com](napatnicky@gmail.com) [thanit.tati@gmail.com](thanit.tati@gmail.com) [schwanph@gmail.com](schwanph@gmail.com)[ ](schwanph@gmail.com)[attapol.t@chula.ac.th](attapol.t@chula.ac.th) [ekapolc@cp.eng.chula.ac.th](ekapolc@cp.eng.chula.ac.th)  \nAbstract  \nLearning job title representation is a vital process for developing automatic human resource tools. To do so, existing methods primarily rely on learning the title representation through skills extracted from the job description, neglecting the rich and diverse content within.  \nThus, we propose an alternative framework for learning job titles through their respective job description (JD) and utilize a Job Description Aggregator component to handle the lengthy description and bidirectional contrastive loss to account for the bidirectional relationship between the job title and its description. We evaluated the performance of our method on both in-domain and out-of-domain settings, achieving a superior performance over the skill-based approach.  \n1 Introduction  \nWith the rapid expansion of online recruitment platforms, vast amounts of job advertisement data (JAD) have been generated. One key part of this data is a job posting, providing detailed information on job titles, specialties, and responsibilities for open positions. Thus, the availability of a system that could understand the post semantics, especially job titles, would greatly facilitate the matchmaking process between both the recruiters and job applicants. This leads to a surge of interest in learning job title representation due to its potential ability to automate job-related tasks such as job recommendation (Kaya and Bogers, 2021 ; Zhao et al., 2021), job trajectory prediction (Decorte et al., 2023a), and job title benchmarking (Zhang et al., 2019) .  \nTo learn the title representation, previous works have primarily relied on utilizing skills information to learn the association between the job title and their respective skill (Decorte et al., 2021 ; Zbibet al., 2022 ; Bocharova et al., 2023) . However, this approach also has some shortcomings as it requires skill information. The skills for a given job are  \neither manually listed, which can be erroneous or incomplete, or automatically extracted from the job description through methods such as keyword matching or automatic skill extraction (Zhang et al., 2022b ; Li et al., 2023) . These skill extraction methods often require a predefined skill vocabulary or acurated dataset (Zhang et al., 2022a) . Furthermore, it is necessary to keep these resources up-to-date with trends in the job market as the dynamic and rapid growth of emerging job roles.  \nPrevious works mitigate these problems by generating synthetic skill data (Decorte et al., 2023b ; Clavié and Soulié, 2023) or creating datasets where both job titles, and skill lists are readily available (Bhola et al., 2020 ; Goyal et al., 2023) . Nonetheless, the former approach further increases pipeline complexity, while the latter suffers from missing skills annotation caused by a communication gap between employers and recruiters.  \nIn this work, we propose to overcome the challenges of obtaining a comprehensive set of skills by bypassing the whole process and instead develop a new framework to learn job titles directly through job descriptions (JDs) without the need for the skill extraction pipeline. We introduce job description aggregation network, which reweights each segment of the JD by their importance and then aggregates th","cbCaigg0HtVFjP9Y","https://ap.wps.com/l/cbCaigg0HtVFjP9Y","pdf",544302,"English","# Abstract\n# Introduction\n# Our Proposed Method\n## Sentence Encoder\n## Job Description Aggregator\n## Contrastive Learning Process","[{\"question\":\"Why is learning job title representation important in recruitment systems?\",\"answer\":\"It enables automation of job-related tasks such as job recommendation, job trajectory prediction, and benchmarking, by helping systems understand post semantics, especially job titles.\"},{\"question\":\"What is the main limitation of prior skill-based approaches?\",\"answer\":\"They depend on skill information, which can be manually listed with errors or be extracted via pipelines that require predefined vocabularies or curated datasets and must be kept up to date.\"},{\"question\":\"How does the proposed method learn job title representations without skill extraction?\",\"answer\":\"It learns from job descriptions by using a Job Description Aggregator network to reweight JD segments into a unified representation, then applies bidirectional contrastive loss to align job titles with their aggregated descriptions.\"}]","Learning Job Title Representation from Job Description Aggregation | PDF"]