Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 136 for “"data annotation"”.

  1. Reducing the effort for data annotation

    … fueled by the increasing availability of large datasets with high-quality annotations. However, the availability of such annotations, also referred to as labels, remains a critical bottleneck for many machine learning applications as they are a prerequisite for supervised model training. This is …

    lmu-germany Repository record for Reducing the effort for data annotation (opens in a new tab)

  2. Addressing the Data Annotation Bottleneck in Breast Digital Pathology

    … with the task performance; however, labeled data is scarce and expensive. The problem is exacerbated in medical image analysis tasks, where an expert must exhaustively annotate data to cover both common and rare cases of disease. In this thesis, I propose techniques to either expedite the …

    toronto-retro Repository record for Addressing the Data Annotation Bottleneck in Breast Digital Pathology (opens in a new tab)

  3. Harnessing Weak Supervision for Reliable and Scalable Data Annotation

    … learning, the need for large-scale labeled datasets has become a critical bottleneck in model development and deployment. Programmatic weak supervision (PWS) offers a promising solution by enabling automatic data labeling through labeling functions (LFs), but faces significant challenges in …

    toronto-retro Repository record for Harnessing Weak Supervision for Reliable and Scalable Data Annotation (opens in a new tab)

  4. Resource-efficient data annotation and augmentation methods for conversational agents

    … taakgerichte chatbots door hun computationele en data-efficiëntie te verbeteren. De thesis introduceert innovatieve methoden voor twee belangrijke taken van chatbots: 1) Automatische data-annotatie: Deze techniek gebruikt unsupervised learning om klantenvragen automatisch in specifieke onderwerpen …

    ghent Repository record for Resource-efficient data annotation and augmentation methods for conversational agents (opens in a new tab)

  5. Streamlining medical data annotation in deep learning through self-supervised approaches

    … continuous increase in available specimens and data is a catalyst for medical research and development and a prerequisite for personalized medicine. One way to effectively process the growing amount of data and derive new diagnostic tools or insights from it is through machine learning …

    fu-berlin Repository record for Streamlining medical data annotation in deep learning through self-supervised approaches (opens in a new tab)

  6. Practical Cost-Conscious Active Learning for Data Annotation in Annotator-Initiated Environments

    … projects exist whose purpose is to augment raw data with annotations that increase the usefulness of the data. The number of these projects is rapidly growing and in the age of “big data” the amount of data to be annotated is likewise growing within each project. One common use of such data is …

    byu Repository record for Practical Cost-Conscious Active Learning for Data Annotation in Annotator-Initiated Environments (opens in a new tab) Download PDF for Practical Cost-Conscious Active Learning for Data Annotation in Annotator-Initiated Environments (opens in a new tab)

  7. Design and implementation of the database module in COLDS for data annotation

    … describes the design and implementation of the Database module in a novel Cloud-based Open Lab for Data Science (COLDS) system. COLDS is a general infrastructure system to support data science programming assignments on the cloud that is currently being developed at the University of Illinois at …

    uiuc Repository record for Design and implementation of the database module in COLDS for data annotation (opens in a new tab)

  8. Data Annotation and Ontology Provisioning for Semantic Applications in Virtualized Wireless Sensor Networks

    … remains as a big challenge. The reason is the data collected by the virtual sensors needs to be annotated in-network, and the pre-requisite of the data annotation process is to have an ontology that needs to be provisioned, i.e., developed, deployed and managed. Unfortunately, annotating sensor …

    concordia Download PDF for Data Annotation and Ontology Provisioning for Semantic Applications in Virtualized Wireless Sensor Networks (opens in a new tab)

  9. Women Workers Behind the AI Revolution: The Production and Reproduction of Data Annotation Platforms

    … work and social relations within transnational data annotation platforms, focusing on the gendered and racialized dynamics in this platform work. The proliferation of the Artificial Intelligence (AI) industry and machine learning companies fundamentally rely upon human labour of data annotation. …

    toronto-retro Repository record for Women Workers Behind the AI Revolution: The Production and Reproduction of Data Annotation Platforms (opens in a new tab)

  10. SARM: Synthetic Data Annotation for Enhancing the Experiences of Augmented Reality Application Based on Machine Learning

    … The research implemented the auto-generated dataset tool, which uses for the Machine Learning dataset preparation step. The final iOS prototype application was developed to incorporate object detection, object tracking and Augmented Reality. The Machine Learning model was taught to recognise …

    auckland-tech Repository record for SARM: Synthetic Data Annotation for Enhancing the Experiences of Augmented Reality Application Based on Machine Learning (opens in a new tab)

  11. How good is good enough? Strategies for dealing with unreliable segmentation annotations of medical data

    … and access to large amounts of medical imaging data with high-quality annotations. However, annotating medical imaging data is time-consuming and expensive, and sometimes the annotations are unreliable.</p> <p>This DPhil thesis presents a comprehensive study that explores deep learning …

    oxford Repository record for How good is good enough? Strategies for dealing with unreliable segmentation annotations of medical data (opens in a new tab)

  12. Three-Dimensional Face Processing and Its Applications in Biometrics

    … of this thesis include: (1) large-scale 3D data collection and data annotation, which resulted in the largest static 3D face database and the first audiovisual 3D dynamic face data collection system in the world; (2) a robust, fast and accurate 3D face reconstruction and animation algorithm, …

    uiuc Repository record for Three-Dimensional Face Processing and Its Applications in Biometrics (opens in a new tab)

  13. Scalable and Generalizable AI Solutions - Applied to Vision-Based Quality Monitoring for Directed Energy Deposition

    … address key challenges in this domain, including data annotation, domain generalization, and real-time adaptability. First, a novel active learning method is proposed to reduce the need for extensive data annotation. This method optimizes model performance by balancing uncertainty with random …

    waterloo Repository record for Scalable and Generalizable AI Solutions - Applied to Vision-Based Quality Monitoring for Directed Energy Deposition (opens in a new tab)

  14. Modeling Comparative Judgments With Machine Learning

    Many tasks can involve analyzing data points and assigning unbound numerical scores or labels. However, without a well-defined standard or point of comparison, such data annotation can require significant cognitive effort from human judges. To mitigate this issue, I propose the use of comparative …

    rit Repository record for Modeling Comparative Judgments With Machine Learning (opens in a new tab)

  15. A Wearable Data Collection Platform with Smart Annotation Capabilities

    … However, to train tracking algorithms, annotated data is needed which is typically obtained manually. Users’ manual annotation is challenging as it’s affected by factors such as recall bias and may become a burden, causing them to stop annotating. Automatic approaches exist, but they may not …

    tamu Repository record for A Wearable Data Collection Platform with Smart Annotation Capabilities (opens in a new tab)

  16. Embracing Annotation Efficient Learning (AEL) for Digital Pathology and Natural Images

    … work effectively. In today's world, the rate of data creation greatly surpasses the rate of data annotation. Full reliance on human annotations is just a temporary means to solve current closed problems in AI. In reality, only a tiny fraction of data is annotated. Annotation Efficient Learning …

    guelph Repository record for Embracing Annotation Efficient Learning (AEL) for Digital Pathology and Natural Images (opens in a new tab)

  17. Regression-Based Active Learning for Scientific Tasks

    Data annotation in some scientific fields, such as materials science and chemistry, can be costly because labels may require complex laboratory experiments or high-fidelity computational simulations. Active learning (AL) as a data selection algorithm aims to reduce the annotation cost by …

    anu Repository record for Regression-Based Active Learning for Scientific Tasks (opens in a new tab)

  18. Exploring Language Models for Legal Applications with Domain Expertise: Interpretability, Evaluation, and Alignment

    … task design. We establish three new tasks and datasets: legal citation prediction, legal hate speech detection, and Misinformation with Legal Consequences (MisLC). Next is the challenge of data annotation; there are significant requirements for data in machine learning (ML) and artificial …

    queens Repository record for Exploring Language Models for Legal Applications with Domain Expertise: Interpretability, Evaluation, and Alignment (opens in a new tab)

  19. A finite-state morphological analyzer for Q'eqchi' using Helsinki Finite-State Technology (HFST) and the Giellatekno infrastructure

    … and sets a foundation for future work in data annotation for this language. The resulting transducer consists of 4,439 lexemes, 2,610 states and 9,558 transitions and covers between 75% and 85% of tokens in a Q'eqchi' corpus. The success of this project lays the groundwork for future work …

    byu Repository record for A finite-state morphological analyzer for Q'eqchi' using Helsinki Finite-State Technology (HFST) and the Giellatekno infrastructure (opens in a new tab) Download PDF for A finite-state morphological analyzer for Q'eqchi' using Helsinki Finite-State Technology (HFST) and the Giellatekno infrastructure (opens in a new tab)

Page 1 of 7