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 30 for “"Model Serving"”.

  1. A large-scale model serving framework

    Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2026-05-01

    uiuc Repository record for A large-scale model serving framework (opens in a new tab)

  2. Multi-tenant Machine Learning Model Serving Systems on GPU Clusters

    … GPUs are both costly and scarce, efficiently serving machine learning models has become a critical challenge. Assuming that serving one model requires $k$ GPUs, serving n models would seemingly require $kn$ GPUs. In the multi-tenant setting, we can pool the whole cluster's GPUs to serve the …

    washington Repository record for Multi-tenant Machine Learning Model Serving Systems on GPU Clusters (opens in a new tab)

  3. Building and Accelerating a Declarative Platform for Machine Learning Model Serving

    … engines to bear for scalable machine learning model serving. Much work remains in AI + systems; advances for each of the subproblems will be applicable to end applications and will lead to better user experiences as well as less cost.

    washington Repository record for Building and Accelerating a Declarative Platform for Machine Learning Model Serving (opens in a new tab)

  4. Assessment of a technical assistance model: Serving youth with disabilities in residential settings in North Dakota.

    … and treatment facilities and school districts in serving these youth.The purpose of this study was to identify if a technical assistance project could facilitate interaction among educators from disparate agencies. The project was evaluated in relation to a model for technical assistance developed …

    nodak Repository record for Assessment of a technical assistance model: Serving youth with disabilities in residential settings in North Dakota. (opens in a new tab)

  5. Retrospective analysis of healthcare utilization and expenditures associated with an integrated, care coordination model serving Medicaid-eligible children with medical complexities

    … effects of a pediatrician-led care coordination model (PLCM) versus usual, standard pediatric care (SPC) on health utilization and expenditure patterns in a randomized cohort of Medicaid-eligible children with medical complexities (CMC). Methods: Retrospective secondary database analysis of Texas …

    texas Repository record for Retrospective analysis of healthcare utilization and expenditures associated with an integrated, care coordination model serving Medicaid-eligible children with medical complexities (opens in a new tab)

  6. Development and Implementation of Control System for an Advanced Multi-Regime Series-Parallel Plug-in Hybrid Electric Vehicle

    Following the Model-Based-Design (MBD) development process used presently by the automotive industry, the control systems for a new Series-Parallel Multiple-Regime Plug-in Hybrid Electric Vehicle (PHEV), UVic EcoCAR2, have been developed, implemented and tested. Concurrent simulation platforms were …

    uvic Repository record for Development and Implementation of Control System for an Advanced Multi-Regime Series-Parallel Plug-in Hybrid Electric Vehicle (opens in a new tab)

  7. Sport Commitment in Wheelchair Basketball: An Interpretive Look into the Lives of Individuals with Physical Disabilities

    … interview method, with the Sport Commitment Model serving as the guiding framework, this study sought to find the underlying factors for continued participation for people with physical disabilities in wheelchair basketball. Through utilizing an interpretive approach it was found that …

    brock Repository record for Sport Commitment in Wheelchair Basketball: An Interpretive Look into the Lives of Individuals with Physical Disabilities (opens in a new tab)

  8. AUTOMATED AND SECURE SYSTEM FOR MACHINE LEARNING LIFECYCLES

    … better predictive performance of downstream models, more efficient transition, more secure remote model serving, etc. Built with the mind of the machine learning pipeline and its evolution, where a pipeline consists of components that flow the raw data to the final model outcome, the system …

    nus Repository record for AUTOMATED AND SECURE SYSTEM FOR MACHINE LEARNING LIFECYCLES (opens in a new tab) Download file for AUTOMATED AND SECURE SYSTEM FOR MACHINE LEARNING LIFECYCLES (opens in a new tab)

  9. Rethinking Serverless for Machine Learning Inference

    … resource utilization. Specifically, we identify model loading and model memory duplication as major bottlenecks in Serverless Inference, and to address these problems, we propose a new approach that rethinks the way we serve FaaS requests. To support this design, we employ a hybrid scaling …

    vt Repository record for Rethinking Serverless for Machine Learning Inference (opens in a new tab)

  10. Resource Sharing for Machine Learning Serving

    … strides in improving machine learning serving efficiency through various approaches, the rapid evolution of machine learning continues to pose new challenges. Traditional design principles, which focus on optimizing individual components such as computation kernels, memory usage, and …

    penn Repository record for Resource Sharing for Machine Learning Serving (opens in a new tab)

  11. Machine Learning-Based Tire Model Optimization

    When developing new vehicles, accurate tire models are crucial for simulating realworld driving behavior. However, inconsistencies in measurement and fitting methods between tire manufacturers introduce bias, leading to different models for the same physical tire. This thesis proposes a method to …

    chalmers Repository record for Machine Learning-Based Tire Model Optimization (opens in a new tab)

  12. Evaluating Incremental Machine Learning Models for Road Condition Classification

    … different incremental Machine Learning (ML) models predicting Road Surface Conditions (RSC), using a given data set from Klimator AB. Successfully identifying the RSC is associated with an autonomous vehicle’s traffic safety and therefore an important area of investigation. This project …

    chalmers Repository record for Evaluating Incremental Machine Learning Models for Road Condition Classification (opens in a new tab)

  13. Research on eeg classification algorithms by sleep stage /

    … a deeper 1D ResNet, and a CNN+BiLSTM sequence model implemented in PyTorch. For per-epoch CNNs we apply a five-state Hidden Markov Model (HMM) to the per-epoch logits as temporal post-processing, and enable the same HMM smoothing in the model serving side to maintain parity with evaluation, for …

    vilnius Repository record for Research on eeg classification algorithms by sleep stage / (opens in a new tab)

  14. Heat Exchanger Design for Waste Heat Recovery using Additive Manufacturing

    … designs. Consequently, this work centers on the modeling, design, optimization, and testing of AM shell and tube heat exchangers for WHR. Specifically, a modified one-dimensional (1+1-D) and three-dimensional (3-D) heat exchanger models were developed and validated, with the 1+1-D model serving …

    ku Repository record for Heat Exchanger Design for Waste Heat Recovery using Additive Manufacturing (opens in a new tab)

  15. InferSuite: A Microarchitectural Characterization and System Profiling of LLM Serving and Agentic Inference - CPU/GPU Time and TMA-Style Breakdown Across Serving and Agentic Workloads

    … use across retrieval, semantic caching, and model serving. Separately from the service, the same measurement method is applied to two existing agentic workloads, a repository-repair agent and a browser agent, driven through their own unmodified harnesses. Everything runs on one bare-metal …

    chalmers Repository record for InferSuite: A Microarchitectural Characterization and System Profiling of LLM Serving and Agentic Inference - CPU/GPU Time and TMA-Style Breakdown Across Serving and Agentic Workloads (opens in a new tab)

  16. Methodologies and Architectures for AI Inference Hardware: From Foundational Networks to Large Language Models

    The rapid advancements in large language models (LLMs) have significantly reshaped the artificial intelligence landscape, enabling transformative applications. However, these developments pose profound challenges for hardware architectures, particularly concerning performance, efficiency, and …

    washington Repository record for Methodologies and Architectures for AI Inference Hardware: From Foundational Networks to Large Language Models (opens in a new tab)

  17. Identifying factors associated with bicycle helmet use behavior among college students

    … was developed based on the Health Belief Model and the Theory of Reasoned Action, with the Stages of Change Model serving as a tool to classify bicycle riders into groups in accordance with current bicycle helmet use behaviors and future intentions to use. A total of 315 responses were …

    colostate Repository record for Identifying factors associated with bicycle helmet use behavior among college students (opens in a new tab)

  18. An empirical investigation of the effects of two goal-setting models on performance and attitudes in a controlled field study

    … to investigate the effects of two goal-setting models in a controlled field study. A review of the literature revealed that two different models might be defined as: (1) the ASPIRATIONAL Model and (2) the ASPIRATIONAL + DIRECTIONAL Model. A treatment condition was based on each of the two …

    utk-theses Repository record for An empirical investigation of the effects of two goal-setting models on performance and attitudes in a controlled field study (opens in a new tab)

  19. Defect Detection for Additive Manufacturing with Machine Learning and Markov Decision Process

    … and develop a Markov Decision Process (MDP) model to make decisions for AM process. Our main purpose is to save costs and time through early termination or parameter adjustment of the printing process.</p> <p>In <em>chapter 1</em>, we developed a scheme based on ML models trained. Then these …

    utk-theses Repository record for Defect Detection for Additive Manufacturing with Machine Learning and Markov Decision Process (opens in a new tab)

Page 1 of 2