[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119689-en":3,"doc-seo-119689-105":30,"detail-sidebar-cat-0-en-105":91},{"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":4,"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},119689,1099513958607,"Jiven","https://ap-avatar.wpscdn.com/avatar/100002390cf8733938c?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778829742770036399",8,"Research & Report","Unified Explanations in Machine Learning Models - A Perturbation Approach","A high-velocity shift toward Explainable Artificial Intelligence (XAI) raises core questions about what complex machine learning models reveal about data and how they reach conclusions. In practice, inconsistencies between explanation methods and modeling techniques can undermine trust in existing explainability. To address this, the work proposes a systematic perturbation-based analysis against SHAP (SHapley Additive exPlanations), offering algorithms for relative feature importance under dynamic inference and measures to quantify how well static explanations remain valid.","Proceedings of the 56th Hawaii International Conference on System Sciences | 2023  \nUnified Explanations in Machine Learning Models: A Perturbation  \nApproach  \nJacob Dineen Arizona State University [j](jdineen@asu.edu)[dineen@asu.edu](jdineen@asu.edu)  \nDon Kridel U Missouri, St Louis [dkridel@gmail.com](dkridel@gmail.com)  \nAbstract  \nA high-velocity paradigm shift towards Explainable Artificial Intelligence (XAI) has emerged in recent years. Highly complex Machine Learning (ML) models have flourished in many tasks of intelligence, and the questions have started to shift away from traditional metrics of validity towards something deeper: What is this model telling me about my data, and how is it arriving at these conclusions? Inconsistencies between XAI and modeling techniques can have the undesirable effect of casting doubt upon the efficacy of these explainability approaches. To address these problems, we propose a systematic, perturbation-based analysis against a popular, model-agnostic method in XAI, SHapley Additive exPlanations (Shap) . We devise algorithms to generate relative feature importance in settings of dynamic inference amongst a suite of popular machine learning and deep learning methods, and metrics that allow us to quantify how well explanations generated under the static case hold. We propose a taxonomy for feature importance methodology, measure alignment, and observe quantifiable similarity amongst explanation models across several datasets.  \n1. Introduction  \nA long-standing problem in predictive analytics has been the disconnect between modelers (statisticians, mathematicians, and data scientists) at the model development stage and end-users at the organizational and decision-maker levels. The latter group is whom the models are ostensibly built for in the first place, but typically does not have the analytical background necessary for a full comprehension of the resulting model artifacts whereas the former community may lack the requisite knowledge of the problem domain driving the requirements of decision-makers. This “cultural divide” too frequently has the undesirable consequence of diminishing the utility of modeling to its intended audience.  \nDaniel Dolk Naval Postgrad School [drdolk@nps.edu](drdolk@nps.edu)  \nDavid Castillo JP Morgan Chase[dcastilloaz@gmail.com](dcastilloaz@gmail.com)  \nThis communication problem has migrated into the arena of machine learning (ML) and artificial intelligence (AI) in recent times, giving rise to the need for and subsequent emergence of Explainable AI (XAI) . XAI has arisen from growing discontent with “blackbox” models, often in the form of neural networksand other emergent, dynamic models (e.g., agent-based simulation, genetic algorithms) that generate outcomes lacking in transparency. This has also been studied through the lens of general machine learning, where classic methods also face an interpretability crisis for high dimensional inputs [1] . Applications such as facial recognition have been met with stern resistance as, too often, mistaken identifications have led to unnecessary and serious disruption in individuals' lives [2, 3, 4] . As another example, the pre-existing bias in historic data-sets used for building ML algorithms (MLAs) has resulted in some people or marginalized communities having firsthand experience with algorithmic unfairness [5, 6] . This is a very sensitive issue for companies for whom the public perception of policy fairness and impartiality is critical to their business and well-being, and has paved the way for increased research interest in algorithmic fairness and equality.  \nA recent well-received book by John Kay and Mervyn King,“Radical Uncertainty: Decision-Making Beyond the Numbers” [7] highlight similar problems for economic models. The authors suggest a need for reference narratives, which are stories that can be marshaled to address the overriding objective of unraveling “what's going on here?” We adopt this perspective as our ","cbCailmcQLFokAWf","https://ap.wps.com/l/cbCailmcQLFokAWf","pdf",587879,1,10,"English","en",105,"# Introduction\n## Motivation and XAI background\n# Contributions\n## Dynamic feature perturbation algorithm\n## Metric derivation\n## Evaluation and comparison\n# Background\n## Preliminaries\n## Static and dynamic scenarios","[{\"question\":\"Why does the paper focus on explainability challenges in machine learning models?\",\"answer\":\"It addresses the gap between explanation goals and modeling realities, where black-box behavior and mismatches between XAI and modeling techniques can reduce trust and interpretability for end-users.\"},{\"question\":\"What is the proposed solution in the paper?\",\"answer\":\"The paper proposes a systematic perturbation-based analysis against SHAP, including algorithms to generate relative feature importance under dynamic inference.\"},{\"question\":\"How does the paper evaluate whether explanations stay reliable over different scenarios?\",\"answer\":\"It derives and adapts distance-based metrics to quantify relative feature importance, then compares similarities between SHAP-based explanations and the proposed dynamic method across datasets.\"}]","Unified Explanations in Machine Learning Models - A Perturbation Approach | PDF",1785725772,25,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":28},"unified-explanations-in-machine-learning-models-a-perturbation-approach","",{"@graph":36,"@context":85},[37,54,68],{"@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/unified-explanations-in-machine-learning-models-a-perturbation-approach/119689/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why does the paper focus on explainability challenges in machine learning models?","Question",{"text":75,"@type":76},"It addresses the gap between explanation goals and modeling realities, where black-box behavior and mismatches between XAI and modeling techniques can reduce trust and interpretability for end-users.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is the proposed solution in the paper?",{"text":80,"@type":76},"The paper proposes a systematic perturbation-based analysis against SHAP, including algorithms to generate relative feature importance under dynamic inference.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the paper evaluate whether explanations stay reliable over different scenarios?",{"text":84,"@type":76},"It derives and adapts distance-based metrics to quantify relative feature importance, then compares similarities between SHAP-based explanations and the proposed dynamic method across datasets.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"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":21,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]