[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-124963-105":59,"doc-detail-124963-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","towards-simplification-of-failure-scenarios-for-machine-learning-enabled-autonomous-systems","Towards Simplification of Failure Scenarios for Machine Learning-enabled Autonomous Systems","","Scenario-based testing is essential for improving the safety and reliability of machine learning-enabled autonomous systems, especially autonomous driving systems. As testing approaches increase realism, failure scenarios become more complex, making it harder to determine root causes. This vision paper proposes simplifying failure scenarios by combining search-based software engineering and surrogate-assisted optimisation to address the key challenges of multiple failure-inducing entity sets, exponential search spaces, limited access to third-party ML components, and costly high-fidelity simulation evaluations.",{"@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/towards-simplification-of-failure-scenarios-for-machine-learning-enabled-autonomous-systems/124963/",{"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/towards-simplification-of-failure-scenarios-for-machine-learning-enabled-autonomous-systems/124963.png","ImageObject",300,407,{"name":92,"@type":93},"Aria","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-09-28","2026-08-05",true,{"@type":102,"interactionType":103,"userInteractionCount":29},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What problem does the paper address for machine learning-enabled autonomous systems?","Question",{"text":112,"@type":113},"It addresses how to simplify a complex failure scenario by identifying a smaller set of scenario entities that still induces the same failure in the MLAS.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"Why is simplifying failure scenarios difficult?",{"text":117,"@type":113},"Failures may be induced by multiple, possibly disjoint entity sets, the candidate search space grows exponentially, third-party ML components restrict access to internal information, and evaluating failures via high-fidelity simulations is computationally expensive.",{"name":119,"@type":110,"acceptedAnswer":120},"How does the proposed vision aim to solve the simplification challenge?",{"text":121,"@type":113},"It combines search-based software engineering with surrogate-assisted optimisation, using surrogate models to reduce the cost of expensive simulation-based fitness evaluations in metaheuristics.","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},124963,1785895655,{"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":29,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":14,"language":139,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":140,"faqs":141,"seo_title":142,"seo_description":67,"update_tm":129,"read_time":24},2336474459895,"https://ap-avatar.wpscdn.com/avatar/22000baeef7a5ed0655?x-image-process=image/resize,m_fixed,w_180,h_180&k=1786071322749376916","Towards Simplification of Failure Scenarios for Machine Learning-enabled Autonomous Systems  \nDonghwan Shin∗ , Sanjeetha Pennada University of Sheffield, Sheffield, United Kingdom  \nd.shin@sheffield.ac.uk, s.pennada@sheffield.ac.uk  \n∗ Corresponding author  \nAbstract—Scenario-based testing is an essential way of improving the safety and reliability of machine learning-enabled autonomous systems (MLAS), such as autonomous driving systems (ADS) . As the complexity of failure scenarios increases with the development of more realistic MLAS testing approaches, it becomes essential to simplify failure scenarios to understand and identify the root causes of failures. In this vision paper, we present our vision to leverage searchbased software engineering (SBSE) and surrogate-assisted optimisation (SAO) to address the challenges of simplifying failure scenarios in MLAS.  \nKeywords–Autonomous Systems, Software Testing, Scenario Simplification  \n1. INTRODUCTION  \nAn autonomous system is an intelligent system (e.g., an autonomous driving system) designed to perceive and acton the surrounding physical environment (e.g., a highway driving environment, including other traffic participants) and work for extended periods of time without human intervention [1] . To ensure the safety and reliability of the autonomous behaviour of the system, therefore, it is essential to consider a scenario. For example, in the automotive domain, a scenario can be defined as a set of static entities (e.g., trees, buildings, traffic signs, and other side objects) and dynamic entities (e.g., moving vehicles, pedestrians, traffic lights, and other state-changing objects) that might affect the behaviour of autonomous driving systems (ADS) . By including all the details (e.g., positions, sizes, shapes, and trajectories) of the static and dynamic entities required to recreate the situation, scenarios become arguably the most important artefacts to verify and validate MLAS [2] .  \nMany approaches have been proposed in the literature to automatically generate failure scenarios with the aim of revealing the failures of MLAS in various domains. One common research direction of these failure scenario generation studies is the increasing complexity of scenarios to capture more realistic situations. For example, early ADS testing studies focused on a few environmental factors (e.g., weather conditions, and road topologies), whereas recent studies start  \nThis is an accepted manuscript, licensed under a Creative Commons Attribution 4 .0 International Public License (CC BY) .  \nThis research is supported by the Engineering and Physical Sciences Research Council [grant number EP/Y014219/1] .  \nto cover more diverse factors (e.g., all controllable factors in a driving simulator with various driving obstacles) in a single scenario. Together with the increasing fidelity of driving simulations [3], the complexity of failure scenarios leads to an essential question: Given a complex failure scenario, what is the minimum set of scenario entities required to cause the same failure?  \nIn this vision paper, we define the problem of MLAS failure scenario simplification with the major challenges to address (Section 2) and propose a research vision to tackle the challenges (Section 3) .  \n2. PROBLEM DEFINITION  \nThe definition of the MLAS failure scenario simplification problem is as follows.  \nDefinition 1 (MLAS Failure Scenario Simplification). For a given MLAS M , its failure scenario S that includes a set of scenario entities, and the failure F of M induced by S, the problem of MLAS failure scenario simplification is to identify a simplified failure scenario S′ that includes the minimum set of scenario entities in S while inducing F.  \nAddressing the problem entails four main challenges. First, due to the non-linear nature of ML models and their interactions in MLAS, a failure scenario might have different (possibly disjoint) sets of entities inducing the same failure. This prevents us","cbCaisAf5hby9Kls","https://ap.wps.com/l/cbCaisAf5hby9Kls","pdf",110177,"English","# Introduction\n## Scenario complexity and the core research question\n# Problem Definition\n## Definition of failure scenario simplification\n## Key challenges\n# Research Vision\n## Combining SBSE and SAO for scenario simplification","[{\"question\":\"What problem does the paper address for machine learning-enabled autonomous systems?\",\"answer\":\"It addresses how to simplify a complex failure scenario by identifying a smaller set of scenario entities that still induces the same failure in the MLAS.\"},{\"question\":\"Why is simplifying failure scenarios difficult?\",\"answer\":\"Failures may be induced by multiple, possibly disjoint entity sets, the candidate search space grows exponentially, third-party ML components restrict access to internal information, and evaluating failures via high-fidelity simulations is computationally expensive.\"},{\"question\":\"How does the proposed vision aim to solve the simplification challenge?\",\"answer\":\"It combines search-based software engineering with surrogate-assisted optimisation, using surrogate models to reduce the cost of expensive simulation-based fitness evaluations in metaheuristics.\"}]","Towards Simplification of Failure Scenarios for Machine Learning-enabled Autonomous Systems | PDF"]