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 46 for “"Privileged Information"”.

  1. Robust Medical Image Analysis using Privileged Information

    … data scarcity and shifts in data distribution. Privileged Information (PI), such as binary attributes or bounding boxes, has the potential to improve machine learning models’ adaptability across diverse domains. This study aims to investigate the role of PI in domain adaptation for medical image …

    chalmers Repository record for Robust Medical Image Analysis using Privileged Information (opens in a new tab)

  2. Domain Adaptation and Privileged Information for Visual Recognition

    … knowledge is better identified by additional information, available during training. Depending on the nature of such information, the learning problem may turn into domain adaptation (DA), domain generalization (DG), leaning using privileged information (LUPI), or domain adaptation with …

    wvu Repository record for Domain Adaptation and Privileged Information for Visual Recognition (opens in a new tab)

  3. Efficient learning with privileged information in nonlinear time series

    … learning algorithms are critical. Learning using privileged information (LuPI) offers increased sample efficiency by allowing prediction models access to information at training time that is unavailable when the models are used. In recent work, it was shown that for prediction in linear- Gaussian …

    chalmers Repository record for Efficient learning with privileged information in nonlinear time series (opens in a new tab)

  4. Metric learning for incorporating privileged information in prototype-based models

    … and novel methodology for incorporating valuable privileged knowledge in the LVQ training phase, but not in testing. This is done by manipulating the global metric in the input space, based on distance relations revealed by the privileged information. Several experiments have been conducted that …

    birmingham Download PDF for Metric learning for incorporating privileged information in prototype-based models (opens in a new tab)

  5. Human activity recognition using conditional random fields and privileged information

    … scheme that is based on the learning using privileged information was also proposed, where privileged information is given as an additional input to the classification model and it is available only during training but never during testing. Experimental results demonstrated that privileged …

    greece Repository record for Human activity recognition using conditional random fields and privileged information (opens in a new tab)

  6. The Role of Sharing and Information Type in Children's Categorization of Privileged and Conventional Information

    … ability to categorize different types of information; specifically conventional information, shareable to others with no restrictions, and privileged information, shareable to only a few. Study 1 investigated 4- and 5-year-olds and adults’ ability to categorize conventional and privileged …

    arkansas Repository record for The Role of Sharing and Information Type in Children's Categorization of Privileged and Conventional Information (opens in a new tab)

  7. Privileged Machine Learning for Prediction

    … from asymmetric distribution, such as posterior information, future information and hidden information. With some additional information only available in training, how to learn a machine learning model with them remains a key challenge. Despite recent advances in important domains such as vision …

    uts Repository record for Privileged Machine Learning for Prediction (opens in a new tab)

  8. Learning using Privileged Time Series

    In this thesis, we study the impact of having privileged information, in the form of intermediate time series, available during the training of models for long-term prediction. An algorithm which incorporates these time series is presented, and we prove that it is more efficient when the time …

    chalmers Repository record for Learning using Privileged Time Series (opens in a new tab)

  9. Identifying the origins of business’ data breaches utilizing covert timing channels

    … businesses. Businesses rely on connectivity and information systems to conduct business, yet those same information systems can be breached and the organization's data exposed. Today, there is a heavy reliance of organizations upon network connections to connect the entire organization in order …

    emich Repository record for Identifying the origins of business’ data breaches utilizing covert timing channels (opens in a new tab)

  10. Learning adaptive dynamics models for robust contact-rich manipulation

    … is trained with an encoded representation of privileged information in an environment with randomized mass and friction conditions. The adaptation module is then trained using a history of observations and actions to match the encoded representation created from the TD-MPC2 algorithm. The …

    texas Repository record for Learning adaptive dynamics models for robust contact-rich manipulation (opens in a new tab)

  11. Classification task-driven efficient feature extraction from tensor data

    … extraction method to derive discriminative privileged information feature from fMRI data. The results from the experiments presented in this thesis demonstrate the advantage of using privileged information for the early detection task.

    birmingham Download PDF for Classification task-driven efficient feature extraction from tensor data (opens in a new tab)

  12. Kernel methods for time series data

    … presents a new methodology for incorporating privileged information about the future evolution of time series, which is only available in the training phase. The task is prediction of the ordered categories of future time series movements. This is implemented by directly extending support …

    birmingham Download PDF for Kernel methods for time series data (opens in a new tab)

  13. Learning Manipulation Policies From Privileged Simulations

    … inputs such as images. Leveraging privileged state when training visuomotor policies significantly reduces the dimensionality of the inputs and the required exploration for reinforcement learning (RL) agents to solve manipulation tasks. My thesis first presents a modular approach …

    cornell Repository record for Learning Manipulation Policies From Privileged Simulations (opens in a new tab)

  14. TOWARDS PRINCIPLED AI AGENTS WITH DECENTRALIZED AND ASYMMETRIC INFORMATION

    … observability, further complicated by asymmetric information between training and testing and decentralized information across agents. This thesis aims to lay the theoretical foundations for principled AI agents operating under partial observability with asymmetric and decentralized information. …

    maryland Repository record for TOWARDS PRINCIPLED AI AGENTS WITH DECENTRALIZED AND ASYMMETRIC INFORMATION (opens in a new tab)

  15. Action Labeling in Images and Video

    … can benefit from being trained with additional information that may be, or may not be, available during deployment. To this end, this dissertation designed, developed, and evaluated methods inspired by the "Learning Using Privileged Information" framework, multimodal data fusion, and knowledge …

    houston Repository record for Action Labeling in Images and Video (opens in a new tab)

  16. Commented translation of excerpts from the novel "Cher Hugo, chère Catherine".

    … this work, the issue of acquiring additional information from the author of the novel is addressed in relation to its effect on the translation; in order to emphasize the need for a carefully balanced translation, this thesis closes by examining the repercussions of abusing this "privileged" …

    ottawa-retro Repository record for Commented translation of excerpts from the novel "Cher Hugo, chère Catherine". (opens in a new tab)

  17. Generalized Boolean functions as combiners

    … Cryptography drives online commerce and allows privileged information safe transit between two parties as well as many other critical internet uses. The presence of a strong pseudo-random number generator (PRNG) is an absolute requirement in modern cryptography. All modern ciphers draw their …

    nps Repository record for Generalized Boolean functions as combiners (opens in a new tab)

  18. Enhancing Materials Discovery Through Advanced Machine Learning Methods

    … a novel ML paradigm called Learning Using Privileged Information (LUPI) for materials' property prediction tasks. This is motivated by the fact that materials characterization through high-throughput simulation not only simulate the target properties, but also generate a wealth of auxiliary …

    umn Repository record for Enhancing Materials Discovery Through Advanced Machine Learning Methods (opens in a new tab)

  19. SAFE REINFORCEMENT LEARNING-BASED GREEN LIGHT OPTIMAL SPEED ADVISORY FOR MIXED-TRAFFIC PLATOONS

    … Vehicles (HVs), we use the concept of privileged information to train the RL agent. Specifically, we leverage both HV and CAV information to train the RL agent while retaining access to only CAV information during implementation. Third, we integrate Control Barrier Functions (CBFs) into …

    nus Repository record for SAFE REINFORCEMENT LEARNING-BASED GREEN LIGHT OPTIMAL SPEED ADVISORY FOR MIXED-TRAFFIC PLATOONS (opens in a new tab) Download file for SAFE REINFORCEMENT LEARNING-BASED GREEN LIGHT OPTIMAL SPEED ADVISORY FOR MIXED-TRAFFIC PLATOONS (opens in a new tab)

  20. Model Based Longitudinal Control of Heavy Duty Vehicles

    … CAN bus, without the need for any manufacturer privileged information. The model-based controller developed was found to be stable and was able to track a wide range of velocities to within 0.5 m/s (~ 1 MPH) of the desired value. Moreover, the plant model developed in Dymola was confirmed to …

    tamu Repository record for Model Based Longitudinal Control of Heavy Duty Vehicles (opens in a new tab)

Page 1 of 3