[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-273649-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-273649-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","reliable-financial-named-entity-recognition-under-domain-shift-confidence-estimation-and-selective-prediction","Reliable Financial Named Entity Recognition under Domain Shift - Confidence Estimation and Selective Prediction","","Financial AI systems that perform named entity recognition often train on one textual register but deploy across SEC filings, financial news, and user-generated social media, where standard F1 scores fail to indicate which extractions remain safe after distribution drift. This work studies confidence estimation and selective prediction for financial NER using calibrated abstention, analyzing a BERT tagger and LoRA-tuned Qwen2.5 models with multiple confidence signals and shift stress tests.",{"@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/reliable-financial-named-entity-recognition-under-domain-shift-confidence-estimation-and-selective-prediction/273649/",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/reliable-financial-named-entity-recognition-under-domain-shift-confidence-estimation-and-selective-prediction/273649.png","ImageObject",300,407,{"name":42,"@type":43},"Riley","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-21","2026-09-15",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",12,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"Why do standard NER metrics like F1 become insufficient under domain shift in financial settings?","Question",{"text":63,"@type":64},"F1 is computed on in-domain test splits and does not reveal which predictions remain trustworthy when inputs move to different registers or topics. Financial NER errors can be harmful because they propagate into compliance, monitoring, and analytical pipelines.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"What models and confidence signals are evaluated for financial NER under shift?",{"text":68,"@type":64},"The study evaluates a BERT encoder tagger and instruction-tuned generative models (Qwen2.5 with LoRA). It compares five inference-time confidence signals, including whole-output probabilities, entity-restricted probabilities, and self-consistency vote share.",{"name":70,"@type":61,"acceptedAnswer":71},"How does abstention affect prediction reliability and coverage across different domains?",{"text":72,"@type":64},"Abstention sharply reduces sentence error on high-confidence in-domain inputs, dropping below 2% on the top 40% highest-confidence fraction. It remains useful on financial news, but does not yield a large clean subset under extreme social-media shift.","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},273649,1789472376,{"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":102,"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":80,"read_time":144,"preview_page_count":4},1374391975076,"https://ap-avatar.wpscdn.com/avatar/14000253ca4ec9f6853?x-image-process=image/resize,m_fixed,w_180,h_180&k=1783305029341752051","Reliable Financial Named Entity Recognition under  \nDomain Shift:  \nConfidence Estimation and Selective Prediction  \nZihao Zheng∗ , Baichuan Li†, Junyi Yao∗ , and Jiayu Long∗  \n∗ Washington University in St. Louis, St. Louis, Missouri, USA  \n†Southern Methodist University, Dallas, Texas, USA  \narXiv : 2608 . 19558v2 [cs .CL] 26 Aug 2026  \nAbstract—Financial AI systems often train information extractors on one textual register and deploy them across filings, news, and user-generated content, and standard F1 scores do not indicate which predictions remain safe to automate when that input distribution changes. We study confidence estimation and selective prediction for financial named entity recognition (NER) on a three-tier stress test spanning SEC filings, financial news, and general-topic social media as an extreme out-ofdomain condition, evaluating a BERT tagger and LoRA-tuned Qwen2.5-0.5B/1.5B models with five inference-time confidence signals, three training seeds, and bootstrap intervals. Confidence rankings themselves change under shift: whole-output probability is the strongest in-domain error detector but deteriorates out of domain, whereas entity-span probability and self-consistency are more robust; self-consistency is also better calibrated without post-hoc fitting. Abstention reduces sentence error from 34.3% to below 2% on the highest-confidence 40% of in-domain inputs and remains useful on financial news, but recovers no usefully large clean subset under the extreme social-media shift. These results motivate a staged deployment strategy that detects severe distribution shift upstream before applying prediction-level confidence gating.  \nIndex Terms—financial named entity recognition, domain shift, confidence estimation, selective prediction, uncertainty estimation, large language models  \nI. INTRODUCTION  \nNamed entity recognition (NER) is a core informationextraction component of applied financial AI systems for compliance, risk, and market-intelligence workflows. A growing line of work fine-tunes small open-weight language models for financial NER with parameter-efficient methods and reports steadily rising F1 on in-domain test splits [1]–[3] . Yet purposebuilt financial LLMs still describe span-level NER as one of their hardest tasks [4]–[6], and a deployed system faces a question that leaderboard F1 does not answer: when the input distribution drifts away from the training data, can the model tell us which of its extractions to trust?  \nThis matters because financial text is not one domain: a model trained on formal filings language (“Apple Inc. reported revenue of $94.9 billion.”) will meet news headlines and social-media commentary (“AAPL just destroyed earnings”) . Prior work shows that neural models become overconfident precisely when inputs shift away from training data [7], [8], while complementary work detects harmful covariate shift at the distribution level before unreliable predictions reach downstream systems [9] . NER models in particular lean on  \nmemorized entity surface forms rather than context [10] . Ina financial setting, a hallucinated organization or a mistyped person entity is not benign: such errors propagate into downstream compliance, monitoring, and analytical pipelines.  \nWe therefore study financial NER through the lens of calibrated selective prediction [11]–[13]: allow the model to abstain on low-confidence predictions and route them to human review, and ask how much reliability this buys at what coverage cost. We fine-tune a BERT encoder tagger and instruction-tuned generative models (Qwen2.5-0.5B and 1.5BInstruct with LoRA [14], [15]) on the FIN corpus of SEC filings [16] and evaluate them on a three-tier stress test: indomain filings (FIN test), financial news (FiNER-ORD [17]), and general-topic social media as an extreme out-of-domain condition (TweetNER7 [18]) . For the generative models we compare five inference-time confidence signals that require no additional training, spa","cbCaid7bhn6DJSAq","https://ap.wps.com/l/cbCaid7bhn6DJSAq","pdf",261554,"English","# Abstract\n# Introduction\n## Problem motivation: reliability under distribution drift\n## Approach: calibrated selective prediction\n## Experimental setup and evaluation\n# Related Work\n## Financial NER datasets and benchmarks\n## Selective prediction and calibration under shift","[{\"question\":\"Why do standard NER metrics like F1 become insufficient under domain shift in financial settings?\",\"answer\":\"F1 is computed on in-domain test splits and does not reveal which predictions remain trustworthy when inputs move to different registers or topics. Financial NER errors can be harmful because they propagate into compliance, monitoring, and analytical pipelines.\"},{\"question\":\"What models and confidence signals are evaluated for financial NER under shift?\",\"answer\":\"The study evaluates a BERT encoder tagger and instruction-tuned generative models (Qwen2.5 with LoRA). It compares five inference-time confidence signals, including whole-output probabilities, entity-restricted probabilities, and self-consistency vote share.\"},{\"question\":\"How does abstention affect prediction reliability and coverage across different domains?\",\"answer\":\"Abstention sharply reduces sentence error on high-confidence in-domain inputs, dropping below 2% on the top 40% highest-confidence fraction. It remains useful on financial news, but does not yield a large clean subset under extreme social-media shift.\"}]","Reliable Financial Named Entity Recognition under Domain Shift - Confidence Estimation and Selective Prediction | PDF",15]