[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-148202-105":59,"doc-detail-148202-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","mind-the-gap-safely-incorporating-deep-learning-models-into-the-actuarial-toolkit","Mind the gap - safely incorporating deep learning models into the actuarial toolkit","","Deep neural network models offer substantial advantages over traditional and machine learning approaches, making them promising for actuarial adoption. Several critical aspects remain insufficiently studied in actuarial literature, including how hyperparameter choices affect prediction accuracy and stability, techniques for generating uncertainty estimates, and design principles for explainable deep learning models. This paper reviews these issues through a deep neural network used to forecast mortality rates, supporting safe toolkit integration.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/mind-the-gap-safely-incorporating-deep-learning-models-into-the-actuarial-toolkit/148202/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/mind-the-gap-safely-incorporating-deep-learning-models-into-the-actuarial-toolkit/148202.png","ImageObject",300,407,{"name":92,"@type":93},"Sage","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-05","2026-08-26",true,{"@type":102,"interactionType":103,"userInteractionCount":44},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"Why are deep neural network models attractive to actuaries?","Question",{"text":112,"@type":113},"The paper explains that deep neural networks can outperform traditional actuarial methods in areas like pricing, reserving, and mortality forecasting using out-of-sample metrics. They also help analyze new high-dimensional or high-frequency data sources.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"What limitations make deep learning risky for actuarial use?",{"text":117,"@type":113},"The paper highlights concerns about “black box” interpretability, the lack of built-in uncertainty measures, and variability across runs due to randomness during training. These factors can be problematic for regulators and other stakeholders.",{"name":119,"@type":110,"acceptedAnswer":120},"How does the paper support safe incorporation of deep learning into actuarial toolkits?",{"text":121,"@type":113},"It reviews key topics needed for safe adoption: effects of hyperparameter selection, methods to produce uncertainty estimates, and approaches for explainable model design. The discussion is framed using a deep neural network for mortality-rate forecasting.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},148202,1787777167,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":44,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":129,"read_time":144,"preview_page_count":4},687197207057,"https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0","British Actuarial Journal (2022), Vol. 27, e21, pp. 1–57 doi:10.1017/S1357321722000162  \nSESSIONAL PAPER  \nMind the gap – safely incorporating deep learning models into the actuarial toolkit  \nRonald Richman  \nOld Mutual Insure, University of the Witwatersrand, Johannesburg, South Africa  \nE-mail: [ronald.richman@ominsure.co.za](ronald.richman@ominsure.co.za), [ronaldrichman@gmail.com](ronaldrichman@gmail.com)  \nAbstract  \nDeep neural network models have substantial advantages over traditional and machine learning methods that make this class of models particularly promising for adoption by actuaries. Nonetheless, several important aspects of these models have not yet been studied in detail in the actuarial literature: the effect of hyperparameter choice on the accuracy and stability of network predictions, methods for producing uncertainty estimates and the design of deep learning models for explainability. To allow actuaries to incorporate deep learning safely into their toolkits, we review these areas in the context of a deep neural network for forecasting mortality rates.  \nKeywords: Deep Learning; Actuarial Science; Model Risk Management; Explainable Machine Learning; Quantile Regression; Mortality Modelling  \n1. Introduction  \nThe maturing field of deep learning, which is the modern approach to designing and fitting neural networks, has achieved remarkable successes in domains such as computer vision, natural language processing and speech recognition (LeCun et al., 2015) and is rapidly advancing the state of the art on so-called “artificial intelligence tasks” (Bengio, 2009), which are those tasks which humans find easy, but have turned out to be difficult for computer systems. These advances have been enabled by the modelling philosophy of representation learning (Bengio et al., 2013; Goodfellow et al., 2016) that is at the core of deep learning, which, in short, can be defined as allowing an algorithm to specify the optimal model structure for a particular task.  \nRecently, the actuarial community has paid increasing attention to deep learning, with investigation of topics across the spectrum of actuarial practice, such as pricing (Noll et al., 2018; Schelldorfer & Wüthrich, 2019), reserving (Gabrielli & Wüthrich, 2018; Gabrielli et al., 2019;Kuo, 2019;Delong et al., 2020), mortality forecasting (Hainaut, 2018; Richman & Wüthrich, 2019;Nigri et al., 2019;Perla et al., 2020;Schnürch&Korn, 2022), life expectancy forecasting (Nigri et al., 2021;Levantesi et al., 2022)and analysis of telematics data (Gao & Wüthrich, 2018; Gao et al., 2019; Gao et al., 2020). Some of these earlier advances have been surveyed in Richman (2018), who provides an introduction to deep learning geared towards an actuarial audience.  \nBased on this emerging area of study within actuarial science, it appears that deep learning is an attractive technique for actuaries for (at least) the following reasons. First, the predictions of deep neural networks have been shown to outperform traditional actuarial techniques in pricing (Richman, 2018; Schelldorfer & Wüthrich, 2019), reserving (Gabrielli et al., 2019) and mortality forecasting (Richman & Wüthrich, 2019), when considering out of sample performance metrics. Second, these techniques enable actuaries to analyse new forms of data (Gao et al., 2020), which are potentially beyond the reach of classical techniques, due to the high dimensionality of the data, the high frequency with which the data are received, or both. Third, actuaries are usually familiar  \n© Institute and Faculty of Actuaries 2022 . This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence ([https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.  \n[https://doi.org/10.1017/S1357321722000162](https://doi.org/10.1017/S1357321722","cbCaicE2gnS9oMrD","https://ap.wps.com/l/cbCaicE2gnS9oMrD","pdf",4695435,57,"English","# Introduction\n## Motivations for deep learning in actuarial work\n## Challenges: black-box behavior and uncertainty\n## Paper objectives and scope","[{\"question\":\"Why are deep neural network models attractive to actuaries?\",\"answer\":\"The paper explains that deep neural networks can outperform traditional actuarial methods in areas like pricing, reserving, and mortality forecasting using out-of-sample metrics. They also help analyze new high-dimensional or high-frequency data sources.\"},{\"question\":\"What limitations make deep learning risky for actuarial use?\",\"answer\":\"The paper highlights concerns about “black box” interpretability, the lack of built-in uncertainty measures, and variability across runs due to randomness during training. These factors can be problematic for regulators and other stakeholders.\"},{\"question\":\"How does the paper support safe incorporation of deep learning into actuarial toolkits?\",\"answer\":\"It reviews key topics needed for safe adoption: effects of hyperparameter selection, methods to produce uncertainty estimates, and approaches for explainable model design. The discussion is framed using a deep neural network for mortality-rate forecasting.\"}]","Mind the gap - safely incorporating deep learning models into the actuarial toolkit | PDF",144]