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 19 of 19 for “"Robust Training"”.
-
Building and using robust representations in image classification
… building learned representations that are more robust and human-aligned. To achieve this, we turn to adversarial (or robust) training, an optimization technique for training networks less prone to adversarial inputs. Typically, robust training is studied purely in the context of machine learning …
-
Evaluating robustness of neural networks with mixed integer programming
… for 4.38% of samples, and a certificate of robustness for the remainder. Across a variety of robust training procedures, we are able to certify more samples than the state-of-the-art and find more adversarial examples than a strong first-order attack for every network.
-
Learning Representations for Limited and Heterogeneous Medical Data
… learning, meta-learning, multitask learning, and robust learning. We present studies with different medical applications, such as clinical language translation, ultrasound image classification and segmentation, medical image retrieval, skin diagnosis classification, pathology metadata prediction, …
-
Statistical Modeling of Knee Morphology and Material Properties Considering Diverse Populations
… further validation with a more balanced and robust training set is needed to confirm these findings. Finally, an innovative application to facilitate the utilization of these findings in the development of orthopedic devices was created. This work represents a significant step towards greater …
-
On the use of email in further education
… for the development of appropriate and robust training to replace current training which is poorly engaged with.
-
Heterogeneous machine learning with decentralized data
… real-world applications, such as collaborative training of financial fraud detection models among banks, or collective health monitoring enabled by massive wearable devices. We investigate three fundamental challenges in this setting. (P1) Effective model training: How can we train models from …
-
Competency requirements of academic librarians in providing research data management services: the case of two university libraries
… research data management services. It recommends robust training as well as institutional support in various forms in order for academic librarians to upskill for effective provision of data management services (RDM) to their research communities. The study also recommends academic library …
-
Cross Target Generalization for Stance Detection
… poorly when exposed to targets unseen during training. Given that data annotation is expensive and time-consuming, finding ways to leverage other sources of knowledge to improve cross-target stance detection can offer great benefits. In this thesis, I suggest to improve the robustness of …
-
Cognitive Radio Engine Design for Link Adaptation
… adaptation. The three areas are CE design, CE training, and the impact of imperfect observations in the operation of the CE. First, we present a CE design for link adaptation and apply it to a system which can adapt its use of multiple antennas in addition to modulation and coding. Our design …
-
Application of foundation models for molecular representation in cancer drug discovery and precision oncology
… of drug repurposing and oncology data, enabling robust training of ML models. We show that these vector embeddings are superior to existing methods, as they enable more accurate drug response predictions. The first part of this thesis is dedicated to development of a deep learning cancer drug …
-
Undergraduate students’ awareness, utilization and satisfaction with the Copperbelt university (CBU) library’s online public access catalogue (OPAC).
… invest in system upgrades, and provide robust training for students. Additionally, continuous monitoring and refining of OPAC implementation based on user feedback is encouraged. Furthermore, future researchers should further investigate demographic influences and conduct comparative …
-
ARTIFICIAL INTELLIGENCE AVAILABLE TO THE DEVELOPMENT OF A VIRTUAL REALITY SOFTWARE FOR AN AUTOMATED CEPHALOMETRIC ANALYSIS OF ULTRA-REDUCED CBCT FOV
… critical for cephalometric analysis. Using a robust training and validation process with expert-annotated CBCT datasets, the system achieved high accuracy, with a mean Euclidean distance error of less than 2 mm. Preprocessing techniques, such as intensity thresholding and data augmentation, …
-
Segmentation and Fracture Detection in CT Images for Traumatic Pelvic Injuries
… that extracts pelvic bone tissues using more robust training models than the Standard ASM algorithm. In addition, a novel hierarchical initialization process for RASM is proposed in order to address the shortcoming of the Standard ASM, i.e. high sensitivity to initialization. Two suitable …
-
Real-time data assimilation in nonlinear dynamical systems
… derive the Jacobian of the network and design a robust training strategy with data augmentation to accurately infer the bias in different scenarios. The r-EnKF is tested on nonlinearly coupled oscillators (with and without time-delay) affected by different forms of bias. The r-EnKF infers in …
-
Structure and thermodynamics of solid/fluid interfaces: simulation and theory
… of becoming trapped in local minima. Using more robust training methods, such as genetic algorithms should be considered for complex force fields with many parameters. It is also demonstrated that even well established reactive force fields that have seen wide adoption in the literature need to …
-
Towards Reliable AI via Efficient Verification of Binarized Neural Networks
… a dearth of prior research that aspires to train robust BNNs. This thesis presents techniques for ensuring neural network robustness against input perturbations and checking safety properties that require a network to produce certain outputs for a set of inputs. We present four contributions: (i) …
-
Speech Foundation Models for Audio Processing
… ASR error correction, we also investigate a training-free, zero-shot approach using large language models (LLMs). This method enables plug-and-play integration with existing ASR systems, without requiring access to model weights or additional training data. Experimental results across …
-
Mental health nursing attitudes towards patient involvement in risk assessment as part of recovery focused care
… was highlighted that there was a lack of training and arguably this lack of understanding could contribute to the lack of value placed on the risk assessment documentation. It is recommended that the NHS trust in the study incorporates more robust training around risk assessment and risk …
-
Robust and high performance machine learning for next generation wireless networks
Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2025-10-19 without embargo terms