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 17 of 17 for “"Feature quality"”.
-
MECHANICAL AND MICROSTRUCTURAL PROPERTIES OF MARAGING STEEL SAMPLES PRODUCED BY ADDITIVE MANUFACTURING (3D PRINTING)
… (TM) in order to compare and contrast part quality. Selective Laser Melting (SLM) is an AM technique which uses a standard triangle language (STL) file to 3D print a part layer by layer. A high-power laser is used to melt metal particles to obtain fully dense parts. This work studied the …
-
A Framework for End-to-End Feature Selection in the Era of Generative AI Agents
… However, growing data volumes and new feature extraction techniques often produce high-dimensional feature spaces containing irrelevant or redundant features. Such features can adversely affect predictive performance, particularly in small-sample settings, where the risk of overfitting …
-
Lightweight Learning-Based Feature Selection for Real Time Optical Flow Navigation
… perform well, their accuracy depends on the quality and stability of visual features. Classical detectors are efficient but often produce redundant or unstable features, while learning-based methods are more robust but computationally expensive for embedded platforms. This thesis presents a …
-
Statistical methods for asset pricing and portfolio management
… a high-dimensional model with noisy predictive features, derives asymptotic expressions for prediction and portfolio performance, and shows that under optimal regularization performance is decreasing and convex in the level of noise. The second studies a partial-observation setting and …
-
Drilling of carbon fibre reinforced plastic (CFRP) and metal matrix composites (MMC)
… MMC composites and understand its effects on feature quality and workpiece integrity. Specific objectives of the project relate to establishing preferred / optimum operating parameters (cutting speed, feed rate and drill strategies) and investigating the influence of cutting environment (dry, …
-
AUTHENTICATING USERS, NOT DEVICES
… Although encryption was achieved, the facial feature extraction software used in this study resulted in poor feature quality. The program seemed to function as designed, but with an insufficient dataset it was difficult to conclude that facial recognition is the best biometric input for …
-
Prior-informed perception and adaptive survey planning for autonomous underwater vehicles
… from multimodal environmental priors, improving feature quality and habitat classification; this benefit is shown to generalise across multiple self-supervised frameworks and architectures, revealing trade-offs between model capacity and generalisation. To support long-duration deployments, an …
-
Development of bioinformatics solutions to enable hair-based exposome research
… hair exposome studies, including optimized feature acquisition, feature quality assurance, intensity correction, top-down annotation, and molecular formula prediction using BUDDY. Out of the 26119 features detected, 316 were annotated using NIST20 and MS-DIAL spectral libraries. BUDDY …
-
Effect of process conditions on feature size of inkjet printed silver MOD ink
… electronics industry necessities uniform, smooth features of high resolution for printed electronic devices. Numerous studies have discussed the impact of print process conditions on the quality of printed features. However, the process conditions of pre-process and post-process techniques …
-
Beyond labels and centralisation: representation learning through data curation
… related and informative batches to improve feature quality under limited and unlabeled data regimes. Second, we introduce FedMPR, a federated parameter-selection framework that adaptively prunes irrelevant weights during local training for each client, improving representation robustness and …
-
Advancing precision medicine through integrative bioinformatics approaches for robust biological knowledge discovery
… medical model that may revolutionize the quality of the current healthcare system. Currently, -omic data are being rapidly accumulated because of the advent of high-throughput -omic assays. Though challenging, abundant information embedded in these data is encouraging for the realization …
-
Scalable optimization-based feature selection with application to recommender systems
… of a variety of data mining techniques, numerous feature selection methods have been introduced to reduce dimensionality. This may improve scalability and make interpreting learning models easier. In this dissertation a new optimization based feature selection method using the nested partition …
-
Robust Algorithms for Estimating Vehicle Movement from Motion Sensors Within Smartphones
… outperformed all the other methods. The feature quality plays a key role in the success of a model. It was found that, the features which exploited the variance of the data were the most effective. In order to assist in quantifying the performance of the machine learning models, a …
-
Electron beam lithography of molecular glass resist films prepared by physical vapor deposition
… importance is thereby the advancement to smaller feature sizes. However, the state of the art exposing technique 193 nm immersion lithography with applied multi-patterning gradually reaches its limits in the ongoing miniaturization. But for the continuation of miniaturization of integrated …
-
Inductive bias from layerwise structure and backpropagation in neural networks: analysis through simplified models and empirical frameworks
… well formalized and understood. We diagonalize features to decouple their learning dynamics and define three properties for each feature: quality, utilization, and intensity. We review the linear model’s inductive bias toward features with larger intensity and demonstrate the benign overfitting …
-
Quantum Spin Hall Effect - A new generation of microstructures
… subsequent lithography steps. With increasing feature quality, more and more physics has been revealed enabling detailed evaluation of smallest disturbances. The following lithography improvements have been implemented. They represent a tool-box for high quality macro and microstructure …
-
Efficient and transferable graph learning via anchor-based structural encodings
… when full GNNs are too expensive or when node features are weak. This dissertation develops and evaluates scalable anchor-based structural encodings as an alternative and complement to conventional message-passing GNN pipelines. The central idea is to represent nodes through their relationships …