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 45 for “"Qwen"”.
-
What Do You Mean? Using Large Language Models for Semantic Evaluation of NL2SQL Queries
… candidate SQL queries through EA, and applies Qwen 2.5 Coder, a 1.5B-parameter model optimized for code generation, to perform semantic equivalence checks. This two-fold process eliminates false positives by detecting logical inconsistencies (e.g., incorrect JOIN conditions or aliases) that EA …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
… (SWIFT) with a powerful vision-language model (Qwen-2.5 VL) to improve species-level predictions on biologically structured datasets. We begin by quantifying the zero-shot capabilities of modern vision language models (VLMs) and demonstrate their limitations in fine-grained settings. We then …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
… (SWIFT) with a powerful vision-language model (Qwen-2.5 VL) to improve species-level predictions on biologically structured datasets. We begin by quantifying the zero-shot capabilities of modern vision language models (VLMs) and demonstrate their limitations in fine-grained settings. We then …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
… (SWIFT) with a powerful vision-language model (Qwen-2.5 VL) to improve species-level predictions on biologically structured datasets. We begin by quantifying the zero-shot capabilities of modern vision language models (VLMs) and demonstrate their limitations in fine-grained settings. We then …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
… (SWIFT) with a powerful vision-language model (Qwen-2.5 VL) to improve species-level predictions on biologically structured datasets. We begin by quantifying the zero-shot capabilities of modern vision language models (VLMs) and demonstrate their limitations in fine-grained settings. We then …
-
Generativní jazykové modely pro automatickou opravu chyb a zvýšení bezpečnosti datových sítí
… jazykovými modely — GPT-OSS 20B, GPT-OSS 120B, Qwen 2.5, Qwen 3.5 a gemma4. Nejlepšího výsledku dosáhla kombinace modelu GPT-OSS 120B v architektuře MAS-, jejíž vážená úspěšnost byla 69,0 % a ostrá úspěšnost 64,3 %. Experimenty ukázaly, že kvalita diagnostiky závisí především na správném výběru …
-
Mašininio mokymosi modelių naudojimas Kubernetes klasterių stebėjimui ir anomalijų aptikimui /
… large-scale language models were compared - Qwen, Meta-Llama and Mistral, of which Qwen achieved the best anomaly detection results. A synthetic dataset of Kubernetes access logs was also generated for anomaly detection experiments. Based on this dataset, an LSTM-based anomaly detection model …
-
MULTILINGUAL EVENT EXTRACTION USING GENERATIVE LARGE LANGUAGE MODELS
… 12 models from four different families (Phi-3, Qwen 2.5, Gemma 2, and Mistral) across three size tiers. The study employs a generative extraction approach using a single prompt across all models to extract four key components from news articles describing protest events: actors, targets, …
-
Towards Green AI: Evaluating Energy Efficiency for Quantized LLM Inference
… on Llama 2 7B, Phi 3.5 Mini Instruct, OLMo 1B, Qwen 2.5 7B, and Qwen 2.5 14B across GLUE MNLI, MMLU, HumanEval, and GSM8K datasets, evaluating 10 post-training quantization (PTQ) strategies on NVIDIA H100, A6000, A100, and L4 GPUs. We identify a novel trend showing that quantization techniques …
-
Multi-Layer OTN Simulation and LLM-Driven Root Cause Analysis: From Alarm Propagation to Reinforcement-Optimized Diagnostic Agents
… combines Supervised Fine-Tuning (SFT) on Qwen 2.5-7B with LoRA adapters and Group Relative Policy Optimization (GRPO) on Qwen 2.5-3B using composite reward functions. A ReAct agent framework wraps the fine-tuned model with five diagnostic tools; a category-based triage layer routes queries …
-
Instruction-Hierarchy Violations and Robustness of Large Language Models to Prompt Injection Attacks
… LLMs, speci cally LLaMA (Meta AI) and unexplored Qwen (Alibaba), which currently rival proprietary systems such as GPT models in reasoning, coding, and multilingual performance. The thesis focuses on simple yet representative attack classes, including Naive Prompt Injection, Fake Completion …
-
Diagnóstico do TEA na era da IA: como assistentes inteligentes podem apoiar a decisão clínica
… artificiais generativas (DeepSeek e Qwen), capaz de identificar possíveis sinais sugestivos de TEA por meio de interações dialógicas orientadas por prompts, gerando relatórios descritivos padronizados para auxiliar profissionais de saúde na triagem e no encaminhamento clínico no …
-
Memory Access Characterization of Large Language Models in CPU Environment and Its Potential Impacts
… two experiments using Llama.cpp and the QWEN model: running various cache configurations and evaluating their performance, and outputting a trace of the memory footprint. Using these experiments, we investigate the memory access patterns and performance characteristics to identify …
-
Explainable Emotion Classification in Social Media
… GPT-4o, LLaMA 3 (8B), and DeepSeek R1 (Distilled Qwen 32B). Building on previous baseline performance, we further improved emotion detection and model interpretability by developing a self-explaining model that uses generative explanations and preference alignment. Notably, we constructed a novel …
-
Conversational Multimodal LLMs for Food Nutritional Information Retrieval: A Systematic Evaluation
… large language models (including GPT-4o, Qwen 2, and Qwen 2.5, DeepSeek, and LLaVA), in a zero-shot setting on two real-world datasets, Nutrition5k and MetaFood3D. To identify which types of information most improve model predictions, a structured cue ladder protocol was developed. Models …
-
Design And Implementation Of LLM-Based Multi-Agent Systems For Automated Data Visualizations
… χρησιμοποιεί το ανοικτού κώδικα μοντέλο Qwen 2.5 7B, το οποίο φιλοξενεί- ται τοπικά για την εξασφάλιση λειτουργικής αυτονομίας και την παράκαμψη των περιορισμών χρήσης των εμπορικών API. Σε αυτή την εκδοχή, ο ρόλος της επικύρωσης ενσωματώνεται απευθείας στον κεντρικό πράκτορα μέσω ενός …
-
LEGAL REASONING WITH SEARCH-AUGMENTED REINFORCEMENT LEARNING: A MULTI-TASK FRAMEWORK
… legal reasoning patterns. Implemented with Qwen 2.5 3B and GRPO optimization on 8 A100 GPUs, the system achieves a 10.5 percentage point improvement over baselines on LegalBench (59.6% vs. 49.7%), 97% jurisdictional compliance, and 0.72 Spearman correlation between model attention and expert …
-
Summarizing Code-Centric Developer Conversations: Analysis and Comparison of Human and LLM-based Approaches
… state-of-the-art LLMs (GPT-4o-mini, DeepSeek-R1, Qwen). Summaries are compared using automated evaluation metrics (ROUGE, BLEU, METEOR, BERTScore) and human evaluation (clarity, conciseness, usefulness, technical accuracy, completeness). GPT-4o-mini and DeepSeek-R1 outperformed other LLMs, and …
-
Systém pro vytěžování zpravodajských informací založený na velkých jazykových modelech
… prohledávání dokumentů, využívající model Qwen 3 32B spolu s modelem BAAI BGE-M3.
-
Daugiaagentė sistema, naudojanti didelės kalbos modelius verslo procesams automatizuoti /
… „Gemma 4 E2B“ till 27 billion parameter „Qwen 3.6 27B“) and two cloud models („GPT‑5.4‑nano“, „GPT‑5.5“) in five scenarios, including a multi‑agent system with extractor, judge, formatter agents. The best local model („Mistral Devstral Small 2 2512“) with „Tesseract“ OCR help achieves the …
Page 1 of 3