Resilience-aware MLOps for AI-based medical diagnostic system
The healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confident...
Uložené v:
| Vydané v: | Frontiers in public health Ročník 12; s. 1342937 |
|---|---|
| Hlavní autori: | , |
| Médium: | Journal Article |
| Jazyk: | English |
| Vydavateľské údaje: |
Switzerland
Frontiers Media S.A
27.03.2024
|
| Predmet: | |
| ISSN: | 2296-2565, 2296-2565 |
| On-line prístup: | Získať plný text |
| Tagy: |
Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
|
| Abstract | The healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confidentiality, bias reduction, model deployment, performance monitoring, and continuous improvement. However, so far, MLOps techniques do not take into account the need to provide resilience to disturbances such as adversarial attacks, including fault injections, and drift, including out-of-distribution. This article is concerned with the MLOps methodology that incorporates the steps necessary to increase the resilience of an AI-based medical diagnostic system against various kinds of disruptive influences.
resilience optimization,
predictive uncertainty calibration, uncertainty monitoring, and graceful degradation are incorporated as additional stages in MLOps. To optimize the resilience of the AI based medical diagnostic system, additional components in the form of adapters and meta-adapters are utilized. These components are fine-tuned during meta-training based on the results of adaptation to synthetic disturbances. Furthermore, an additional model is introduced for
calibration of predictive uncertainty. This model is trained using both in-distribution and out-of-distribution data to refine predictive confidence during the inference mode.
The structure of resilience-aware MLOps for medical diagnostic systems has been proposed. Experimentally confirmed increase of robustness and speed of adaptation for medical image recognition system during several intervals of the system's life cycle due to the use of resilience optimization and uncertainty calibration stages. The experiments were performed on the DermaMNIST dataset, BloodMNIST and PathMNIST. ResNet-18 as a representative of convolutional networks and MedViT-T as a representative of visual transformers are considered. It is worth noting that transformers exhibited lower resilience than convolutional networks, although this observation may be attributed to potential imperfections in the architecture of adapters and meta-adapters.
The main novelty of the suggested resilience-aware MLOps methodology and structure lie in the separating possibilities and activities on creating a basic model for normal operating conditions and ensuring its resilience and trustworthiness. This is significant for the medical applications as the developer of the basic model should devote more time to comprehending medical field and the diagnostic task at hand, rather than specializing in system resilience. Resilience optimization increases robustness to disturbances and speed of adaptation. Calibrated confidences ensure the recognition of a portion of unabsorbed disturbances to mitigate their impact, thereby enhancing trustworthiness. |
|---|---|
| AbstractList | The healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confidentiality, bias reduction, model deployment, performance monitoring, and continuous improvement. However, so far, MLOps techniques do not take into account the need to provide resilience to disturbances such as adversarial attacks, including fault injections, and drift, including out-of-distribution. This article is concerned with the MLOps methodology that incorporates the steps necessary to increase the resilience of an AI-based medical diagnostic system against various kinds of disruptive influences.BackgroundThe healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confidentiality, bias reduction, model deployment, performance monitoring, and continuous improvement. However, so far, MLOps techniques do not take into account the need to provide resilience to disturbances such as adversarial attacks, including fault injections, and drift, including out-of-distribution. This article is concerned with the MLOps methodology that incorporates the steps necessary to increase the resilience of an AI-based medical diagnostic system against various kinds of disruptive influences.Post-hoc resilience optimization, post-hoc predictive uncertainty calibration, uncertainty monitoring, and graceful degradation are incorporated as additional stages in MLOps. To optimize the resilience of the AI based medical diagnostic system, additional components in the form of adapters and meta-adapters are utilized. These components are fine-tuned during meta-training based on the results of adaptation to synthetic disturbances. Furthermore, an additional model is introduced for post-hoc calibration of predictive uncertainty. This model is trained using both in-distribution and out-of-distribution data to refine predictive confidence during the inference mode.MethodsPost-hoc resilience optimization, post-hoc predictive uncertainty calibration, uncertainty monitoring, and graceful degradation are incorporated as additional stages in MLOps. To optimize the resilience of the AI based medical diagnostic system, additional components in the form of adapters and meta-adapters are utilized. These components are fine-tuned during meta-training based on the results of adaptation to synthetic disturbances. Furthermore, an additional model is introduced for post-hoc calibration of predictive uncertainty. This model is trained using both in-distribution and out-of-distribution data to refine predictive confidence during the inference mode.The structure of resilience-aware MLOps for medical diagnostic systems has been proposed. Experimentally confirmed increase of robustness and speed of adaptation for medical image recognition system during several intervals of the system's life cycle due to the use of resilience optimization and uncertainty calibration stages. The experiments were performed on the DermaMNIST dataset, BloodMNIST and PathMNIST. ResNet-18 as a representative of convolutional networks and MedViT-T as a representative of visual transformers are considered. It is worth noting that transformers exhibited lower resilience than convolutional networks, although this observation may be attributed to potential imperfections in the architecture of adapters and meta-adapters.ResultsThe structure of resilience-aware MLOps for medical diagnostic systems has been proposed. Experimentally confirmed increase of robustness and speed of adaptation for medical image recognition system during several intervals of the system's life cycle due to the use of resilience optimization and uncertainty calibration stages. The experiments were performed on the DermaMNIST dataset, BloodMNIST and PathMNIST. ResNet-18 as a representative of convolutional networks and MedViT-T as a representative of visual transformers are considered. It is worth noting that transformers exhibited lower resilience than convolutional networks, although this observation may be attributed to potential imperfections in the architecture of adapters and meta-adapters.The main novelty of the suggested resilience-aware MLOps methodology and structure lie in the separating possibilities and activities on creating a basic model for normal operating conditions and ensuring its resilience and trustworthiness. This is significant for the medical applications as the developer of the basic model should devote more time to comprehending medical field and the diagnostic task at hand, rather than specializing in system resilience. Resilience optimization increases robustness to disturbances and speed of adaptation. Calibrated confidences ensure the recognition of a portion of unabsorbed disturbances to mitigate their impact, thereby enhancing trustworthiness.СonclusionThe main novelty of the suggested resilience-aware MLOps methodology and structure lie in the separating possibilities and activities on creating a basic model for normal operating conditions and ensuring its resilience and trustworthiness. This is significant for the medical applications as the developer of the basic model should devote more time to comprehending medical field and the diagnostic task at hand, rather than specializing in system resilience. Resilience optimization increases robustness to disturbances and speed of adaptation. Calibrated confidences ensure the recognition of a portion of unabsorbed disturbances to mitigate their impact, thereby enhancing trustworthiness. BackgroundThe healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confidentiality, bias reduction, model deployment, performance monitoring, and continuous improvement. However, so far, MLOps techniques do not take into account the need to provide resilience to disturbances such as adversarial attacks, including fault injections, and drift, including out-of-distribution. This article is concerned with the MLOps methodology that incorporates the steps necessary to increase the resilience of an AI-based medical diagnostic system against various kinds of disruptive influences.MethodsPost-hoc resilience optimization, post-hoc predictive uncertainty calibration, uncertainty monitoring, and graceful degradation are incorporated as additional stages in MLOps. To optimize the resilience of the AI based medical diagnostic system, additional components in the form of adapters and meta-adapters are utilized. These components are fine-tuned during meta-training based on the results of adaptation to synthetic disturbances. Furthermore, an additional model is introduced for post-hoc calibration of predictive uncertainty. This model is trained using both in-distribution and out-of-distribution data to refine predictive confidence during the inference mode.ResultsThe structure of resilience-aware MLOps for medical diagnostic systems has been proposed. Experimentally confirmed increase of robustness and speed of adaptation for medical image recognition system during several intervals of the system’s life cycle due to the use of resilience optimization and uncertainty calibration stages. The experiments were performed on the DermaMNIST dataset, BloodMNIST and PathMNIST. ResNet-18 as a representative of convolutional networks and MedViT-T as a representative of visual transformers are considered. It is worth noting that transformers exhibited lower resilience than convolutional networks, although this observation may be attributed to potential imperfections in the architecture of adapters and meta-adapters.СonclusionThe main novelty of the suggested resilience-aware MLOps methodology and structure lie in the separating possibilities and activities on creating a basic model for normal operating conditions and ensuring its resilience and trustworthiness. This is significant for the medical applications as the developer of the basic model should devote more time to comprehending medical field and the diagnostic task at hand, rather than specializing in system resilience. Resilience optimization increases robustness to disturbances and speed of adaptation. Calibrated confidences ensure the recognition of a portion of unabsorbed disturbances to mitigate their impact, thereby enhancing trustworthiness. The healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confidentiality, bias reduction, model deployment, performance monitoring, and continuous improvement. However, so far, MLOps techniques do not take into account the need to provide resilience to disturbances such as adversarial attacks, including fault injections, and drift, including out-of-distribution. This article is concerned with the MLOps methodology that incorporates the steps necessary to increase the resilience of an AI-based medical diagnostic system against various kinds of disruptive influences. resilience optimization, predictive uncertainty calibration, uncertainty monitoring, and graceful degradation are incorporated as additional stages in MLOps. To optimize the resilience of the AI based medical diagnostic system, additional components in the form of adapters and meta-adapters are utilized. These components are fine-tuned during meta-training based on the results of adaptation to synthetic disturbances. Furthermore, an additional model is introduced for calibration of predictive uncertainty. This model is trained using both in-distribution and out-of-distribution data to refine predictive confidence during the inference mode. The structure of resilience-aware MLOps for medical diagnostic systems has been proposed. Experimentally confirmed increase of robustness and speed of adaptation for medical image recognition system during several intervals of the system's life cycle due to the use of resilience optimization and uncertainty calibration stages. The experiments were performed on the DermaMNIST dataset, BloodMNIST and PathMNIST. ResNet-18 as a representative of convolutional networks and MedViT-T as a representative of visual transformers are considered. It is worth noting that transformers exhibited lower resilience than convolutional networks, although this observation may be attributed to potential imperfections in the architecture of adapters and meta-adapters. The main novelty of the suggested resilience-aware MLOps methodology and structure lie in the separating possibilities and activities on creating a basic model for normal operating conditions and ensuring its resilience and trustworthiness. This is significant for the medical applications as the developer of the basic model should devote more time to comprehending medical field and the diagnostic task at hand, rather than specializing in system resilience. Resilience optimization increases robustness to disturbances and speed of adaptation. Calibrated confidences ensure the recognition of a portion of unabsorbed disturbances to mitigate their impact, thereby enhancing trustworthiness. |
| Author | Kharchenko, Vyacheslav Moskalenko, Viacheslav |
| AuthorAffiliation | 2 Department of Computer Systems, Network and Cybersecurity, Faculty of Radio-Electronics, Computer Systems and Infocommunications, National Aerospace University “KhAI” , Kharkiv , Ukraine 1 Department of Computer Science, Faculty of Electronics and Information Technologies, Sumy State University , Sumy , Ukraine |
| AuthorAffiliation_xml | – name: 1 Department of Computer Science, Faculty of Electronics and Information Technologies, Sumy State University , Sumy , Ukraine – name: 2 Department of Computer Systems, Network and Cybersecurity, Faculty of Radio-Electronics, Computer Systems and Infocommunications, National Aerospace University “KhAI” , Kharkiv , Ukraine |
| Author_xml | – sequence: 1 givenname: Viacheslav surname: Moskalenko fullname: Moskalenko, Viacheslav – sequence: 2 givenname: Vyacheslav surname: Kharchenko fullname: Kharchenko, Vyacheslav |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/38601490$$D View this record in MEDLINE/PubMed |
| BookMark | eNp9UU1v1DAQtVARLaV_gAPKkUuW8Uec-IBQVfGx0qJKCM7WxBlvXWXjxc6C-u_xdpeq5cDBsjV-896beS_ZyRQnYuw1h4WUnXnnt7v-ZiFAqAWXShjZPmNnQhhdi0Y3J4_ep-wi51sA4CAVCP6CncpOA1cGztj7b5TDGGhyVONvTFR9XV1vc-Vjqi6XdY-ZhmpDQ3A4VkPA9RTzHFyV7_JMm1fsuccx08XxPmc_Pn38fvWlXl1_Xl5drmqntJlrbIqu68XeLRqS2HaSNCH0GjpfpmndoAdJAOi86XrXdL4zispBpUjIc7Y88A4Rb-02hQ2mOxsx2PtCTGuLqdgayUottOBOSnBKNdCj1uA9DJ4MiUFC4fpw4CoLLIM5muaE4xPSpz9TuLHr-MtyDqCE1IXh7ZEhxZ87yrPdhOxoHHGiuMu2iLTStK2SBfrmsdiDyt8ACkAcAC7FnBP5BwgHuw_a3gdt96uzx6BLU_dPkwszziHuDYfxf61_AE3orYg |
| CitedBy_id | crossref_primary_10_1007_s00330_025_11654_6 crossref_primary_10_3390_info16020087 crossref_primary_10_1109_ACCESS_2025_3606838 crossref_primary_10_3390_s24185951 crossref_primary_10_2196_66559 crossref_primary_10_1002_pros_24871 crossref_primary_10_1016_j_engappai_2024_109649 crossref_primary_10_3390_app15052628 |
| Cites_doi | 10.1109/comst.2020.3036778 10.3390/app12199851 10.1109/tii.2022.3149935 10.1109/tkde.2022.3178128 10.1007/s10994-023-06336-7 10.1016/B978-0-12-824020-5.00028-4 10.1016/j.media.2021.102141 10.54517/m.v4i1.2156 10.3389/frai.2022.950659 10.32620/reks.2020.3.05 10.1007/s11219-022-09601-5 10.1016/j.compbiomed.2023.106848 10.1109/jsait.2020.2991430 10.1371/journal.pone.0265723 10.1007/978-3-030-40245-7_15 10.1007/978-3-030-86365-4_54 10.1016/j.artmed.2023.102718 10.1109/tnnls.2022.3183120 10.3390/app12031353 10.3390/a16030165 10.30844/wi_2020_c1-baier 10.15588/1607-3274-2023-2-9 10.3233/jifs-179677 10.1038/s42256-023-00626-4 10.1016/j.cose.2021.102367 10.1109/tai.2022.3159510 10.32620/reks.2020.1.02 10.1109/tpami.2022.3191696 10.1016/j.compbiomed.2023.106791 10.1148/radiol.2020200038 10.2478/s13537-011-0025-x 10.1109/access.2022.3181730 |
| ContentType | Journal Article |
| Copyright | Copyright © 2024 Moskalenko and Kharchenko. Copyright © 2024 Moskalenko and Kharchenko. 2024 Moskalenko and Kharchenko |
| Copyright_xml | – notice: Copyright © 2024 Moskalenko and Kharchenko. – notice: Copyright © 2024 Moskalenko and Kharchenko. 2024 Moskalenko and Kharchenko |
| DBID | AAYXX CITATION CGR CUY CVF ECM EIF NPM 7X8 5PM DOA |
| DOI | 10.3389/fpubh.2024.1342937 |
| DatabaseName | CrossRef Medline MEDLINE MEDLINE (Ovid) MEDLINE MEDLINE PubMed MEDLINE - Academic PubMed Central (Full Participant titles) DOAJ Directory of Open Access Journals |
| DatabaseTitle | CrossRef MEDLINE Medline Complete MEDLINE with Full Text PubMed MEDLINE (Ovid) MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic MEDLINE |
| Database_xml | – sequence: 1 dbid: DOA name: DOAJ Directory of Open Access Journals url: https://www.doaj.org/ sourceTypes: Open Website – sequence: 2 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Public Health |
| EISSN | 2296-2565 |
| ExternalDocumentID | oai_doaj_org_article_362621c330c4450ba660ff0dfe9e2d30 PMC11004236 38601490 10_3389_fpubh_2024_1342937 |
| Genre | Journal Article |
| GroupedDBID | 53G 5VS 9T4 AAFWJ AAYXX ACGFO ACGFS ADBBV ADRAZ AFPKN ALMA_UNASSIGNED_HOLDINGS AOIJS BAWUL BCNDV CITATION DIK EMOBN GROUPED_DOAJ GX1 HYE KQ8 M48 M~E OK1 RNS RPM ACXDI CGR CUY CVF ECM EIF IPNFZ NPM RIG 7X8 5PM |
| ID | FETCH-LOGICAL-c469t-a5402cb22024a9e3a783e6ea0b608f3897cd6d3e00acf98bc58f894e894a44e23 |
| IEDL.DBID | DOA |
| ISICitedReferencesCount | 9 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001198920600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 2296-2565 |
| IngestDate | Fri Oct 03 12:41:22 EDT 2025 Thu Aug 21 18:34:28 EDT 2025 Fri Sep 05 06:04:42 EDT 2025 Mon Jul 21 05:55:53 EDT 2025 Sat Nov 29 04:16:39 EST 2025 Tue Nov 18 22:35:19 EST 2025 |
| IsDoiOpenAccess | true |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Keywords | medical diagnosis robustness MLOps resilience image recognition |
| Language | English |
| License | Copyright © 2024 Moskalenko and Kharchenko. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c469t-a5402cb22024a9e3a783e6ea0b608f3897cd6d3e00acf98bc58f894e894a44e23 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 John Plodinec, Independent Researcher, Aiken, SC, United States Edited by: Tetyana Chumachenko, Kharkiv National Medical University, Ukraine Reviewed by: Gilbert Yong San Lim, SingHealth, Singapore |
| OpenAccessLink | https://doaj.org/article/362621c330c4450ba660ff0dfe9e2d30 |
| PMID | 38601490 |
| PQID | 3037397743 |
| PQPubID | 23479 |
| ParticipantIDs | doaj_primary_oai_doaj_org_article_362621c330c4450ba660ff0dfe9e2d30 pubmedcentral_primary_oai_pubmedcentral_nih_gov_11004236 proquest_miscellaneous_3037397743 pubmed_primary_38601490 crossref_primary_10_3389_fpubh_2024_1342937 crossref_citationtrail_10_3389_fpubh_2024_1342937 |
| PublicationCentury | 2000 |
| PublicationDate | 2024-03-27 |
| PublicationDateYYYYMMDD | 2024-03-27 |
| PublicationDate_xml | – month: 03 year: 2024 text: 2024-03-27 day: 27 |
| PublicationDecade | 2020 |
| PublicationPlace | Switzerland |
| PublicationPlace_xml | – name: Switzerland |
| PublicationTitle | Frontiers in public health |
| PublicationTitleAlternate | Front Public Health |
| PublicationYear | 2024 |
| Publisher | Frontiers Media S.A |
| Publisher_xml | – name: Frontiers Media S.A |
| References | Huang (ref24) 2020; 1 Li (ref34) 2020 Wang (ref21) 2022; 35 Gupta (ref4) 2019 Duddu (ref14) 2020; 38 Stirbu (ref2) 2022; 31 Olowononi (ref13) 2021; 23 Niemelä (ref5) 2022 Khattak (ref6) 2023 Xu (ref10) 2020 Islam (ref25) 2022; 12 Zhang (ref12) 2022 Vassilev (ref36) 2023 Nirmala (ref17) 2022; 5 Bortsova (ref18) 2021; 73 Lusenko (ref49) 2020; 1 Qiu (ref33) 2020 Abusnaina (ref22) 2021 Ponochovnyi (ref48) 2020; 3 Pianykh (ref30) 2020; 297 Testi (ref1) 2022; 10 Kalin (ref9) 2022 Shen (ref39) 2022 Xu (ref35) 2019 Moskalenko (ref16) 2023; 16 Moskalenko (ref41) 2023; 2 Dymond (ref46) 2021 Silva Filho (ref44) 2023; 112 Kotyan (ref45) 2022; 17 Liu (ref31) 2023; 147 Song (ref50) 2020 Chen (ref38) 2023 Li (ref42) 2018 Jiao (ref32) 2024; 35 Hou (ref40) 2021 Ding (ref37) 2023; 5 Pourpanah (ref26) 2022; 45 Yang (ref11) 2021; 12893 Subramanya (ref3) 2022; 12 Manzari (ref51) 2023; 157 Peng (ref43) 2023; 19 Khalid (ref7) 2023; 158 Guelfi (ref47) 2011; 1 Fang (ref8) 2023; 4 Achddou (ref28) 2021 Baier (ref27) 2020 Shapeev (ref29) 2020; 968 Duy (ref15) 2021; 109 Karimi (ref23) 2023; 4 Awais (ref19) 2021 Gongye (ref20) 2020 |
| References_xml | – volume: 23 start-page: 524 year: 2021 ident: ref13 article-title: Resilient machine learning for networked cyber physical systems: a survey for machine learning security to securing machine learning for CPS publication-title: IEEE Commun Surv Tutor doi: 10.1109/comst.2020.3036778 – volume: 12 start-page: 9851 year: 2022 ident: ref3 article-title: From dev ops to MLOps: overview and application to electricity market forecasting publication-title: Appl Sci doi: 10.3390/app12199851 – year: 2022 ident: ref39 – volume: 19 start-page: 2463 year: 2023 ident: ref43 article-title: Open-set fault diagnosis via supervised contrastive learning with negative out-of-distribution data augmentation publication-title: IEEE Trans Industr Inform doi: 10.1109/tii.2022.3149935 – volume: 35 start-page: 1 year: 2022 ident: ref21 article-title: Generalizing to unseen domains: a survey on domain generalization publication-title: IEEE Trans Knowl Data Eng doi: 10.1109/tkde.2022.3178128 – volume: 112 start-page: 3211 year: 2023 ident: ref44 article-title: Classifier calibration: a survey on how to assess and improve predicted class probabilities publication-title: Mach Learn doi: 10.1007/s10994-023-06336-7 – year: 2019 ident: ref35 – year: 2021 ident: ref40 – start-page: 183 volume-title: Adversarial Robustness for Machine Learning year: 2023 ident: ref38 article-title: Adversarial robustness in meta-learning and contrastive learning doi: 10.1016/B978-0-12-824020-5.00028-4 – volume: 73 start-page: 102141 year: 2021 ident: ref18 article-title: Adversarial attack vulnerability of medical image analysis systems: unexplored factors publication-title: Med Image Anal doi: 10.1016/j.media.2021.102141 – volume: 4 start-page: 17 year: 2023 ident: ref8 article-title: A survey on adversarial attack and defense of deep learning models for medical image recognition publication-title: Meta doi: 10.54517/m.v4i1.2156 – volume: 5 start-page: 950659 year: 2022 ident: ref17 article-title: A novel technique for detecting sudden concept drift in healthcare data using multi-linear artificial intelligence techniques publication-title: Front Artific Intellig doi: 10.3389/frai.2022.950659 – volume: 3 start-page: 43 year: 2020 ident: ref48 article-title: Dependability assurance methodology of information and control systems using multipurpose service strategies publication-title: Radioelectron Comput Syst doi: 10.32620/reks.2020.3.05 – year: 2021 ident: ref46 – volume: 31 start-page: 307 year: 2022 ident: ref2 article-title: Continuous design control for machine learning in certified medical systems publication-title: Softw Qual J doi: 10.1007/s11219-022-09601-5 – volume: 158 start-page: 106848 year: 2023 ident: ref7 article-title: Privacy-preserving artificial intelligence in healthcare: techniques and applications publication-title: Comput Biol Med doi: 10.1016/j.compbiomed.2023.106848 – volume: 1 start-page: 267 year: 2020 ident: ref24 article-title: Functional error correction for robust neural networks publication-title: IEEE J Select Areas Info Theory doi: 10.1109/jsait.2020.2991430 – volume: 17 start-page: e0265723 year: 2022 ident: ref45 article-title: Adversarial robustness assessment: why in evaluation both L0 and L∞ attacks are necessary publication-title: PLoS One doi: 10.1371/journal.pone.0265723 – year: 2020 ident: ref10 – volume: 968 start-page: 309 year: 2020 ident: ref29 article-title: Active learning and uncertainty estimation publication-title: Mach Learn Meets Quant Phys doi: 10.1007/978-3-030-40245-7_15 – volume: 12893 start-page: 671 year: 2021 ident: ref11 article-title: Uncertainty quantification and estimation in medical image classification publication-title: Lect Notes Comput Sci doi: 10.1007/978-3-030-86365-4_54 – year: 2021 ident: ref22 – volume: 147 start-page: 102718 year: 2023 ident: ref31 article-title: A few-shot disease diagnosis decision making model based on meta-learning for general practice publication-title: Artif Intell Med doi: 10.1016/j.artmed.2023.102718 – volume: 35 start-page: 1278 year: 2024 ident: ref32 article-title: Dynamic ensemble selection for imbalanced data streams with concept drift publication-title: IEEE Trans Neural Netw Learn Syst doi: 10.1109/tnnls.2022.3183120 – year: 2020 ident: ref34 – volume: 12 start-page: 1353 year: 2022 ident: ref25 article-title: A systematic review of explainable artificial intelligence in terms of different application domains and tasks publication-title: Appl Sci doi: 10.3390/app12031353 – year: 2018 ident: ref42 – year: 2020 ident: ref20 – year: 2023 ident: ref6 – volume: 16 start-page: 165 year: 2023 ident: ref16 article-title: Resilience and resilient systems of artificial intelligence: taxonomy, models and methods publication-title: Algorithms doi: 10.3390/a16030165 – year: 2021 ident: ref19 – start-page: 210 volume-title: WI2020 Zentrale Tracks. year: 2020 ident: ref27 article-title: Handling concept drifts in regression problems—the error intersection approach doi: 10.30844/wi_2020_c1-baier – year: 2020 ident: ref33 – year: 2021 ident: ref28 – year: 2020 ident: ref50 – volume: 2 start-page: 79 year: 2023 ident: ref41 article-title: Model-agnostic Meta-learning for resilience optimization of artificial intelligence system publication-title: Radio Electron Comput Sci Control doi: 10.15588/1607-3274-2023-2-9 – year: 2022 ident: ref5 – volume: 38 start-page: 5897 year: 2020 ident: ref14 article-title: Fault tolerance of neural networks in adversarial settings publication-title: IFS doi: 10.3233/jifs-179677 – year: 2022 ident: ref12 – volume: 5 start-page: 220 year: 2023 ident: ref37 article-title: Parameter-efficient fine-tuning of large-scale pre-trained language models publication-title: Nat Mach Intellig doi: 10.1038/s42256-023-00626-4 – volume: 109 start-page: 102367 year: 2021 ident: ref15 article-title: DIGFuPAS: deceive IDS with GAN and function-preserving on adversarial samples in SDN-enabled networks publication-title: Comput Secur doi: 10.1016/j.cose.2021.102367 – volume: 4 start-page: 383 year: 2023 ident: ref23 article-title: Improving calibration and out-of-distribution detection in deep models for medical image segmentation publication-title: IEEE Trans Artific Intellig doi: 10.1109/tai.2022.3159510 – year: 2023 ident: ref36 – year: 2019 ident: ref4 – volume: 1 start-page: 17 year: 2020 ident: ref49 article-title: Computer systems resilience in the presence of cyber threats: Taxonomy and ontology publication-title: Radioelectron Comput Syst doi: 10.32620/reks.2020.1.02 – volume: 45 start-page: 1 year: 2022 ident: ref26 article-title: A review of generalized zero-shot learning methods publication-title: IEEE Trans Pattern Anal Mach Intell doi: 10.1109/tpami.2022.3191696 – year: 2022 ident: ref9 – volume: 157 start-page: 106791 year: 2023 ident: ref51 article-title: Med ViT: a robust vision transformer for generalized medical image classification publication-title: Comput Biol Med doi: 10.1016/j.compbiomed.2023.106791 – volume: 297 start-page: 6 year: 2020 ident: ref30 article-title: Continuous learning AI in radiology: implementation principles and early applications publication-title: Radiology doi: 10.1148/radiol.2020200038 – volume: 1 start-page: 294 year: 2011 ident: ref47 article-title: A formal framework for dependability and resilience from a software engineering perspective publication-title: Open Comput Sci doi: 10.2478/s13537-011-0025-x – volume: 10 start-page: 63606 year: 2022 ident: ref1 article-title: MLOps: a taxonomy and a methodology publication-title: IEEE Access doi: 10.1109/access.2022.3181730 |
| SSID | ssj0001034021 |
| Score | 2.31567 |
| Snippet | The healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems.... BackgroundThe healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI)... |
| SourceID | doaj pubmedcentral proquest pubmed crossref |
| SourceType | Open Website Open Access Repository Aggregation Database Index Database Enrichment Source |
| StartPage | 1342937 |
| SubjectTerms | Artificial Intelligence Awareness Data Accuracy image recognition Machine Learning medical diagnosis MLOps Public Health resilience Resilience, Psychological robustness |
| Title | Resilience-aware MLOps for AI-based medical diagnostic system |
| URI | https://www.ncbi.nlm.nih.gov/pubmed/38601490 https://www.proquest.com/docview/3037397743 https://pubmed.ncbi.nlm.nih.gov/PMC11004236 https://doaj.org/article/362621c330c4450ba660ff0dfe9e2d30 |
| Volume | 12 |
| WOSCitedRecordID | wos001198920600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVAON databaseName: DOAJ Directory of Open Access Journals customDbUrl: eissn: 2296-2565 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0001034021 issn: 2296-2565 databaseCode: DOA dateStart: 20130101 isFulltext: true titleUrlDefault: https://www.doaj.org/ providerName: Directory of Open Access Journals – providerCode: PRVHPJ databaseName: ROAD: Directory of Open Access Scholarly Resources customDbUrl: eissn: 2296-2565 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0001034021 issn: 2296-2565 databaseCode: M~E dateStart: 20130101 isFulltext: true titleUrlDefault: https://road.issn.org providerName: ISSN International Centre |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV1Lb9QwEB5BxQEJId4sj8pI3JCpY3sd-8ChoFYgtQUhQHuzHHsiVipptbuFG7-dsZ2udhGCC4fkkDjJeGacmbFnPgM8V60N5OhKntAprpNV3OooudShpwGILZaqtC9H7cmJnc3ch42tvnJOWIUHrozby3ApsokUdketp6ILxoi-F6lHhzKpEq2T17MRTJXZFaEoMGpqlQxFYW4vl1vlxQepXzaKfsJ54_MNS1QA-__kZf6eLLlhfQ5vwc3RbWT7ldzbcAWHO3CjzrmxWkp0F159xOX8tIxVHn6EBbLjo_fnS0Z-Kdt_x7PBSuxbXZlhqebY0ftYRXO-B58PDz69ecvH7RF4pJh2xQM5WzJ2MvcpOFShtQoNBtEZYXvqcxuTSQqFCLF3totT21unkY6gNUp1H3aGswEfArPShY56KE3oddJT50ISIlHbJqYozASaS1b5OGKH5y0sTj3FEJm9vrDXZ1L8yN4JvFg_c16RM_7a-nWWwLplRr0uF0gX_KgL_l-6MIFnl_LzNEry0kcY8Oxi6clQt8XVVRN4UOW5_pSyJseJ9LTdkvQWLdt3hvnXgsRd8PakMo_-B_WP4XrmSE5wk-0T2FktLvApXIvfV_PlYheutjO7W7Sczsc_D34B5Z8AZg |
| linkProvider | Directory of Open Access Journals |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Resilience-aware+MLOps+for+AI-based+medical+diagnostic+system&rft.jtitle=Frontiers+in+public+health&rft.au=Moskalenko%2C+Viacheslav&rft.au=Kharchenko%2C+Vyacheslav&rft.date=2024-03-27&rft.issn=2296-2565&rft.eissn=2296-2565&rft.volume=12&rft.spage=1342937&rft_id=info:doi/10.3389%2Ffpubh.2024.1342937&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2296-2565&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2296-2565&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2296-2565&client=summon |