[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117293-en":3,"doc-seo-117293-105":29,"detail-sidebar-cat-0-en-105":95},{"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":20,"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":13,"seo_description":14,"update_tm":27,"read_time":28},117293,8796095461564,"Liam","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","The Coevolution of Human and Machine Learning - Thesis Abstract","Machine learning has achieved remarkable progress, yet it faces major crises: heavy reliance on large annotated datasets, degraded performance when training with insufficient or low-quality data, and limited human interpretability due to black-box behavior. These issues hinder human trust and restrict practical adoption. To enable joint progress of human and machine learning, domain knowledge is used to improve accuracy under limited data, while interactive multimodel collaboration with experts strengthens performance and trustworthiness under low-quality data. An explainable multimodal framework further clarifies outputs using privileged text and annotations during training, improving applicability across domains.","The Coevolution of Human and  \nMachine learning by Siqi Zhang  \nThesis submitted in fulfilment of the requirements for the degree of  \nDoctor of Philosophy  \nunder the supervision of Prof. Yang Wang, A/Prof. Zhidong Li and Prof. Richard Xu  \nUniversity of Technology Sydney  \nFaculty of Engineering and Information Technology 06/2023  \nCERTIFICATE OF ORIGINAL AUTHORSHIP  \nI, Siqi Zhang, declare that this thesis is submiCed in fulfilment of the requirements for the award of Doctor of Philosophy, in the Faculty of Engineering and informaKon Technology at the University of Technology Sydney.  \nThis thesis is wholly my own work unless otherwise referenced or acknowledged. In addiKon, I cerKfy that all informaKon sources and literature used are indicated in the thesis.  \nThis document has not been submiCed for qualificaKons at any other academic insKtuKon.  \nThis research is supported by the Australian Government Research Training Program.  \nSignature: Production Note:  \nSignature removed prior to publication.  \nDate:05/30/2023  \nPage 1 of 1 Cer*ficate of Original Authorship, last updated 07/07/2022  \n2  \nABSTRACT  \nMachine learning has made remarkable progress in the past decade, and its application scope and depth continue to expand. However, it also faces significant crises and challenges. Firstly, although machine learning has high accuracy for model output, they rely heavily on large amounts of annotated data. In situations with limited data or no annotations, ensuring the accuracy for the model output becomes a significant goal for specific industries. Secondly, machine learning applications often encounter scenarios with insufficient data or low quality data. When such data is used to train models, it can lead to significant deviations, resulting in inaccurate output and a decline in people’s trust in machine learning. Thirdly, many machine learning models operate in a black box environment, where the models often do not provide explanations or the explanations provided are too complex. Without appropriate feedback, humans cannot understand the learning status of the model and cannot effectively intervene in the model’s learning. Therefore, it is difficult for machine learning to gain human trust. This issue can have a direct impact on the application and advancement of machine learning.  \nTo address these challenges, under the guidance of my supervisors, I conducted research on the coevolution of human and machine learning. Our research goal is to enable human and machine learning models to progress together, achieving better performance, gaining people’s understanding and trust ,and then applying it. To improve the accuracy of output results in situations with limited data, we introduced domain knowledge to jointly train the model. For scenarios with lowquality data, we proposed a multi-model structure that fosters interaction and collaboration between models and experts. This approach allows experts to monitor and enhance model performance, boosting the trustworthiness of machine  \n3  \nlearning models. To address situations where there is no explanation or unreasonable explanation, we developed an explainable machine learning framework that uses multimodal methods to clarify the output results of machine learning models for non-machine learning experts. This framework promotes the broader application of machine learning. Our specific work is as follows:  \n1. We proposed a Bayesian Nonparametric Process(BNP) method for adding rule-based domain knowledge, enabling effective training even with limited and unannotated training data. The utility function of domain knowledge was integrated into the model as a prior, forming a responsive BNP method that can quickly learn from input data and achieve excellent performance in limited data situations. We validated the proposed method on both a simulated dataset anda supermarket dataset, achieving outstanding results.  \n2. We proposed a framework for interactive collaboration between mode","cbCaibODb6sBAhLy","https://ap.wps.com/l/cbCaibODb6sBAhLy","pdf",3026740,1,121,"English","en",105,"# Abstract\n## Key challenges and motivations\n## Proposed coevolution approach\n## BN P domain-knowledge method\n## Interactive model-expert collaboration framework\n## Explainable multimodal information framework","[{\"question\":\"Why does machine learning struggle with trust and adoption?\",\"answer\":\"Models often depend on large annotated datasets, suffer when data is limited or low quality, and behave as black boxes that provide no clear or usable explanations. Without appropriate feedback, humans cannot intervene effectively.\"},{\"question\":\"How does the research improve learning when data is limited and unannotated?\",\"answer\":\"It introduces a Bayesian Nonparametric Process method that integrates rule-based domain knowledge as a prior utility, enabling responsive training from input data with strong performance under limited-data conditions.\"},{\"question\":\"What approach is used to handle low-quality or insufficient data?\",\"answer\":\"A multimodel interactive collaboration framework is proposed, where experts validate and guide student models during learning. Additional expert domain knowledge further steers training and improves trustworthiness.\"},{\"question\":\"How is explainability provided for non-machine-learning experts?\",\"answer\":\"An explainable multimodal information framework uses privileged information such as text and expert annotations during training to clarify model outputs, while avoiding the need for these inputs during testing for broader usability.\"}]",1785675040,305,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":90,"head_meta":92,"extra_data":94,"updated_unix":27},"the-coevolution-of-human-and-machine-learning-thesis-abstract","",{"@graph":35,"@context":89},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/the-coevolution-of-human-and-machine-learning-thesis-abstract/117293/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-08-05","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"Why does machine learning struggle with trust and adoption?","Question",{"text":75,"@type":76},"Models often depend on large annotated datasets, suffer when data is limited or low quality, and behave as black boxes that provide no clear or usable explanations. Without appropriate feedback, humans cannot intervene effectively.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the research improve learning when data is limited and unannotated?",{"text":80,"@type":76},"It introduces a Bayesian Nonparametric Process method that integrates rule-based domain knowledge as a prior utility, enabling responsive training from input data with strong performance under limited-data conditions.",{"name":82,"@type":73,"acceptedAnswer":83},"What approach is used to handle low-quality or insufficient data?",{"text":84,"@type":76},"A multimodel interactive collaboration framework is proposed, where experts validate and guide student models during learning. Additional expert domain knowledge further steers training and improves trustworthiness.",{"name":86,"@type":73,"acceptedAnswer":87},"How is explainability provided for non-machine-learning experts?",{"text":88,"@type":76},"An explainable multimodal information framework uses privileged information such as text and expert annotations during training to clarify model outputs, while avoiding the need for these inputs during testing for broader usability.","https://schema.org",{"og:url":51,"og:type":91,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":93,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,119,124,127,132,135,139],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},6,"Technology",50,"technology",{"id":120,"doc_module":4,"doc_module_name":45,"category_name":121,"show_sort_weight":122,"slug":123},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":125,"slug":126},30,"research-report",{"id":128,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":130,"slug":131},9,"Religion & Spirituality",20,"religion-spirituality",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":130,"slug":134},"World Cup","world-cup",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":136,"slug":138},10,"Lifestyle","lifestyle",{"id":140,"doc_module":4,"doc_module_name":45,"category_name":141,"show_sort_weight":110,"slug":142},19,"General","general"]