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 16 of 16 for “"BERT-based models"”.
-
Comparative Study of Deep Learning Models for Sentiment Analysis
… study of three CNN variants and seven BERT-based models for sentiment analysis. BERT variants were achieved from pre-trained "bert-base-uncased" and "roberta-case" models, hand-crafted prompt, adaptive prompt, and hybrid prompt. The Hybrid GRU_ATT BERT model outperforms all others. …
-
Semantic Message Representations for Behavioral Prediction in a Diabetes Prevention System
… message content is represented limits what these models can learn. This thesis studies a diabetes prevention system that sends weekly health messages to participants in rural India and predicts their responses. The original system encodes each message as a binary identifier, treating every piece …
-
Preclinical risk of bias assessment and PICO extraction using natural language processing
… thesis, I explore natural language processing models for facilitating two stages in preclinical systematic reviews: risk of bias assessment and PICO extraction. There are a range of measures used to reduce bias in animal experiments and many checklist criteria require the reporting of those …
-
Understanding stance classification of BERT models : an attention-based mechanism
BERT produces state-of-the-art solutions for many natural language processing tasks at the cost of interpretability. As works discuss the value of BERT’s attention weights to this purpose, we contribute with an attention-based interpretability framework to identify the most influential words for …
-
Natural language processing for business documents
… unstructured public data and deploying language models cost-effectively, especially for resource-limited organizations like small-to-medium enterprises (SMEs). Addressing these challenges is crucial for applications ranging from fraud detection (Goel and Gangolly, 2012), long-form summarization …
-
Enhancing Generative AI-Based Health Misinformation Detection with Debunk Lists as an External Knowledge Framework
… misinformation, traditional machine learning models often require large, labeled datasets for fine-tuning, whereas generative AI models, particularly large language models (LLMs), offer a promising alternative due to their strong generalization capabilities. This dissertation examines whether …
-
Data-Driven Analysis of Crane Accident Reports Using Structured and Narrative Information
… D. Wiethorn, P.E., PhD: damage level estimation based on structured crane accident information and responsible personnel classification using narrative descriptions. Across Experiments, the work examines how heterogeneous data contributes to model performance and interpretability. For damage …
-
Προηγμένες τεχνικές μηχανικής μάθησης για πρόβλεψη και ταξινόμηση με χρήση συναισθηματικής ανάλυσης και ενεργής μάθησης
… sentiment configurations (TextBlob, VADER, FinBERT). The findings highlight the positive contribution of Sentiment Analysis to improving economic forecasts. Subsequently, the study focuses on Financial Sentiment Analysis as a classification problem, comparing on the one hand numerous classical …
-
Improving Data Quality in an Institutional Clinical Trial Data Repository to Support Patient-Trial Matching
… Encoder Representations from Transformers (BERT) model, has shown promise in many NLP tasks and been the basis for other BERT-based models pre-trained with domain-specific resources. My thesis aimed to evaluate the abilities of biomedical-domain-specific BERT models to discriminate between …
-
Steering textual reasoning with explanations
… extended the boundaries of language models' (LMs') potential applications, partially because of their increased ability to do complex reasoning. However, LMs have well-documented reasoning failures, such as hallucinations and inability to systematically generalize. In this …
-
Crisis detection from Arabic social media
… amounts of unlabelled text. Results show that BERT-based models are more robust to out-of-distribution target events and remarkably outperform other models on the information classification task. Experiments show that the performance of BERT-based classifiers can be enhanced when training on …
-
Robust and Interpretable Cross-Domain Arabic Readability Prediction Using Hybrid Modeling: Evidence on the Limitations of Transformer-Based Classifiers
… of statistical classifiers and transformer-based models, including CAMeLBERT(Inoue et al, 2021.) with a focus on their ability to generalize under domain, genre, and stylistic shifts. Second, I investigate the interpretability of transformer performance in readability task, which are often …
-
Automatic Essay Scoring in Arabic: Development, Evaluation, and Advanced Techniques
… with improvements of up to 40% observed using BERT-based models for English essays. <br/><br/>Given the limited resources for Arabic AES, this thesis introduces the AR-AES dataset, consisting of 2046 essays from undergraduate students across multiple courses, annotated independently by two …
-
Data curation of a findable, accessible, interoperable, reusable polymer nanocomposites data resource - MaterialsMine
… ChemProps, a RESTful API-enabled multi-algorithm-based polymer/filler name mapping methodology, are discussed in detail in this dissertation. To create truly powerful and transformative materials design paradigms and towards a sustainable future for MaterialsMine, we need to harness the power of …
-
Using natural language processing to address health professionals’ information needs
… providing direct answers to clinicians’ queries based on relevant, reliable, and up-to-date information. However, despite advances in biomedical QA, adoption in clinical settings remains limited.<br/><br/>This thesis aims to investigate the barriers to adopting biomedical QA systems and to …
-
Development of machine-learning-based natural language processing to detect concept labels in clinical narratives
… important insights into using machine learning models in limited datasets. Specifically, we found that smaller and simpler models can work better in such contexts. To this end, our framework combines TF-IDF and MLP-NN, and we demonstrate that feature selection from the learning representation …