[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121842-en":3,"doc-seo-121842-105":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},121842,1649267921044,"Ava Thompson","https://us-avatar.wpscdn.com/avatar/1800007509477c92dfb?_k=1782875107921204101",8,"Research & Report","Problems and Solutions - Machine Learning Approaches for a Dynamic Ocean","The dissertation presents machine learning solutions for a dynamic ocean by addressing the challenges of dataset shift and domain adaptation across changing conditions. It focuses on joint learning of classifiers and background weighting, robust hard-negative selection, and image preprocessing and augmentation to improve generalization. The work evaluates multiple adaptation and fusion strategies with quantitative testing results, including accuracy and F1-score metrics, and includes qualitative comparisons and ablation studies to identify the most effective components.","UC San Diego  \nUC San Diego Electronic Theses and Dissertations  \nTitle  \n# Problems and Solutions:Machine Learning Approaches for a Dynamic Ocean\n\nPermalink  \nhttps://escholarship.org/uc/item/70j2r271  \nAuthor  \nWalker,Joseph Leslie  \nPublication Date  \n2023  \nPeer reviewed|Thesis/dissertation  \nUNIVERSITY OF CALIFORNIA SAN DIEGO  \nProblems and Solutions:Machine Learning Approaches for a Dynamic Ocean  \nA Dissertation submitted in partial satisfaction of the requirementsfor the degree Doctor of Philosophy  \nin  \nOceanography  \nby  \nJoseph Leslie Walker  \nCommittee in charge:  \nKaitlin Frasier,ChairFlorian MeyerStuart SandinNuno Vasconcelos  \nCopyright  \nJoseph Leslie Walker,2023  \nAll rights reserved.  \nThe Dissertation of Joseph Leslie Walker is approved,and it isacceptable in quality and form for publication on microfilm andelectronically.  \nUniversity of California San Diego  \n2023  \nTABLE OF CONTENTS  \nDISSERTATION APPROVALPAGE……………………………………………………………………………………………ii  \nTABLE OF CONTENTS………………………………………………………………………………………………………………………iv  \nLIST OF FIGURES………………………………………………………………………………………………………………………………V  \nLIST OF TABLES………………………………………………………………………………………………………………………………………………vii  \nACKNOWLEDGEMENTS………………………………………………………………………………………………………………viii  \nVITA……………………………………………………………………………………………………………………………………………………………X  \nABSTRACT OF THE DISSERTATION…………………………………………………………………………………………………xXi  \nChapter 1 INTRODUCTION………………………………………………………………………………………………………………………1  \nChapter ……………………………………………………………………………………………………………………………………………………6  \nAcknowledgements ………………………………………………………………………………………………………………………………………………16  \nChapter 3…………………………………………………………………………………………………………………………………………………………………17  \nAcknowledgements ………………………………………………………………………………………………………………………………………………44  \nChapter 4…………………………………………………………………………………………………………………………………………………………………45  \nAcknowledgements ………………………………………………………………………………………………………………………………………72  \n## LIST OF FIGURES\n\nFigure 1:Using target and background training datasets,the parameters of a classifierand background image weights are jointly learned using an alternative optimizationapproach.Testing is then performed using target,novel,and dataset shiftdatasets…  ……………………  \nFigure 2:Three examples,selected by a human annotator,from classes in Dtarg (leftcolumn)and Dout (right column)showing the morphological similarity betweenspecimen of these classes………      …   ……  \n9  \nFigure 3:Class distribution of the downsampled background sets generated from eachof the downsampling strategies.a)class distribution for the 15 least abundant classesthat have at least 100 examples in the original Dout set.b)class distribution for thefive most abundant classes in Dout.……                  …     10  \nFigure 4:Samples from each target class and their respective hard negatives.Tenbackground classes are represented by the hard negativeexamples……                ……                 11  \nFigure 5:Daily average background image weight value associated with each of thehard background classes presented in the order:a)Dictyocha;b)detritus;c)Skeletonema;d)pennate.The bars represent the standard deviation of the weightvalues…                     …       ……………  …                      …   …      11  \nFigure 6:0OD testing results.Left-to-right:Accuracy and F1 score obtained from anaverage of model runs using ratios of{1:400,1:50,1:10,1:1}.Error bars reflectstandard deviation…              ……      …  12  \nFigure 7:Sampling locations,dates,methodology,and a SUIT……………………………………    24  \nFigure 8:Example images collected from the six study sites.The bounding boxescontaining the SUITs are shown in orange……………………………………………     27  \nFigure 9:Data augmentations.a)An example image from the TAK study site.Images  \nb)-f)show the output of the translation,rotation,perspective transformation,cropping,and distance image augmentation functions respectively using the image in a)as input.    29  \nFigure 10:a)Overview of our augmented version of the progressive domainadaptation method,using the HM-MVG","cbCaikqrA99TWtDU","https://ap.wps.com/l/cbCaikqrA99TWtDU","pdf",34578877,1,85,"English","en",105,"# Chapter 1 INTRODUCTION\n## List of Figures\n## List of Tables\n## Acknowledgements\n## VITA\n# ABSTRACT OF THE DISSERTATION\n# Chapter 3\n# Chapter 4","[{\"question\":\"What is the central problem addressed by this dissertation?\",\"answer\":\"How to apply machine learning to a dynamic ocean setting while overcoming dataset shift between training and target conditions.\"},{\"question\":\"Which major methods are developed to improve learning robustness?\",\"answer\":\"The dissertation includes joint learning of classifier and background weights, hard-negative selection, and domain adaptation enhanced by augmentation and adversarial feature alignment.\"},{\"question\":\"How are the approaches evaluated?\",\"answer\":\"Models are tested using quantitative metrics such as accuracy and F1-score, supported by dataset-shift testing results, qualitative image comparisons, and ablation studies.\"}]","Problems and Solutions - Machine Learning Approaches for a Dynamic Ocean | PDF",1785807178,214,{"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},"problems-and-solutions-machine-learning-approaches-for-a-dynamic-ocean","",{"@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/problems-and-solutions-machine-learning-approaches-for-a-dynamic-ocean/121842/",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","2026-08-04",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},"What is the central problem addressed by this dissertation?","Question",{"text":75,"@type":76},"How to apply machine learning to a dynamic ocean setting while overcoming dataset shift between training and target conditions.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which major methods are developed to improve learning robustness?",{"text":80,"@type":76},"The dissertation includes joint learning of classifier and background weights, hard-negative selection, and domain adaptation enhanced by augmentation and adversarial feature alignment.",{"name":82,"@type":73,"acceptedAnswer":83},"How are the approaches evaluated?",{"text":84,"@type":76},"Models are tested using quantitative metrics such as accuracy and F1-score, supported by dataset-shift testing results, qualitative image comparisons, and ablation studies.","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},[93,97,101,105,110,115,120,123,128,131,135],{"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":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]