18736 research outputs found
Sort by
Enhancing News Article Generation with AI Tools
Get PDFAbstract. This paper aims to present a comparative analysis of a custom-built AI Journalist Assistant and ChatGPT 4.0 for news article generation. The goal is to evaluate the performance of each model based on accuracy, speed, ethical safeguards, and relevance, particularly in the context of journalism. While ChatGPT is widely used for general-purpose content creation, its reliance on older data and potential for plagiarism presents challenges in the fast-paced, high-stakes world of news reporting. To address these issues, we will design and implement an AI Assistant using Retrieval-Augmented Generation (RAG) techniques, focusing on real-time data access, bias reduction, and plagiarism prevention. By comparing both models through established metrics such as BERTScore, precision, recall, and inference time, this paper contributes insights into the role of specialized AI tools in journalistic workflows and highlights the trade-offs between prompt engineering and domain-specific AI development. The findings will guide future advancements in AI-generated news and its ethical implications
Advancing Statistical Methods for Multivariate and Network Meta-Analysis
Get PDFMultivariate meta-analysis (MMA) and network meta-analysis (NMA) are essential tools for synthesizing evidence across multiple correlated outcomes and treatments. However, these tools face practical challenges, including outcome reporting bias (ORB), unreported within-study correlations, and computational burden. ORB can distort effect estimates in MMA, while missing within-study correlations in multivariate NMA may lead to biased conclusions. To address these challenges, this dissertation introduces two novel statistical methods. For MMA, we propose SemiMMA, a semiparametric and scalable approach that treats ORB as a missing-not-at-random problem and combines inverse propensity weighting (IPW) with the generalized method of moments (GMM). For multivariate NMA, we develop BCL-OFS, a calibrated Bayesian composite likelihood method that employs a hybrid Gibbs sampler and Open-Faced Sandwich adjustment to accommodate unreported within-study correlations without requiring a fully specified likelihood. Through extensive simulations and real-world applications, including studies of treatments for kidney disease and root coverage procedures, we demonstrate that SemiMMA yields robust estimates against ORB, while BCL-OFS produces unbiased estimates with nominal levels. These methodological advancements mitigate key biases and offer the reliability of meta-analytic conclusions in medical and public health research
Leveraging GenAI for Biometric Voice Print Authentication
Get PDFThis paper presents the development of a secure voice authentication system that delivers an inclusive solution for all users, including those with disabilities. Leveraging a Text-Dependent Active Verification process, the system combines a spoken passphrase with voice biometric coefficients and audio vector embeddings for reliable user verification. A vector database is used to efficiently store data and perform similarity retrieval. Initially, the system achieves a 71% spoof detection accuracy, ensuring that only genuine samples proceed to the embedding stage, where it attains a 55.21% accuracy in vector embedding and similarity retrieval. Furthermore, this approach paves the way for user-specific voice-controlled environments. Overall, this study underscores the transformative potential of voice biometric authentication by merging cutting-edge signal processing with sophisticated machine learning techniques, setting a benchmark for future research in balancing robust security measures, user convenience, and ethical inclusivity
Destruction of Castle Savoia (Castle Savoy) with Dynamic Weather Effects
Get PDFThis thesis project focused on three masteries: Historical Recreation, Static Destruction, and Dynamic Weather. The Castle Savoia is recreated in the original and the destroyed version, to be switched between different weather types. Each weather type would affect the scene dynamically, either making the castle appear wet or covered by accumulated snow
CFO Narcissism and the Power of Persuasion Over Analysts: A Mixed-Methods Approach
No full textWe study the role of CFO narcissism in the intent and ability to positively influence sell-side analysts’ perceptions of the firm. Consistent with narcissists casting favorable impressions on others, we find CFO narcissism is associated with overly optimistic analyst valuations. We then study public persuasion attempts by analyzing conference call transcripts and private persuasion attempts through a laboratory study. In the conference call setting, we show that narcissistic CFOs use more persuasive language and are more inclined to call on bearish analysts, both of which we link to price target revisions following the call. In the lab study, we simulate a one-on-one conversation and find that narcissists are especially more likely to use coercive methods to induce higher valuations from analysts. Collectively, we show that narcissistic CFOs use persuasion to favorably influence analysts’ perceptions of firm value
Implementing Defamiliarization Techniques to Enhance Quest-based Storytelling
Get PDFDefamiliarization is a technique originating from literary theory that can be used to subvert player expectations and create novel and thought-provoking experiences. This thesis explores how players react to defamiliarization and how defamiliarization enhances players’ narrative experiences. The researcher designed two levels in Fallout 4. One level contained no unusual or unexpected elements, serving as a baseline, while the other level built on the same structure as the first level, but incorporated five selected defamiliarization techniques. By analysing player reactions and feedback, the researcher gained insights into how game designers can effectively leverage defamiliarization to craft more engaging and immersive storytelling experiences
Mitochondrial Stress-Induced Carboxyl-Terminal Extensions: A Critical Link Between Ribosome Associated Quality Control and Mitochondrial Quality Control
No full textRibosome-associated quality control (RQC) plays a crucial role in cellular homeostasis by identifying and resolving stalled translation events. Increasing evidence suggests a link between RQC and mitophagy, the selective degradation of damaged mitochondria. In this study, we demonstrate that blocking msiCTE inhibits mitophagy, establishing it as a key regulatory factor in this process. We further show that mitochondrial stress recruits RQC factors, facilitating selective mitophagy through specific autophagy receptors. Additionally, our findings reveal that a certain length of artificial CAT-tails can trigger mitophagy and that modulating RQC factors directly impacts mitophagy flux. These insights deepen our understanding of how ribosome surveillance mechanisms influence mitochondrial quality control and cellular stress adaptation
Examining Working and Episodic Memory in Young Adults with Anhedonia
Get PDFDepression is associated with impairments in memory processes. Evidence suggests there is poor recognition of positive information and quicker disengagement with positive information from working memory in depression compared to healthy controls. The working memory differences are hypothesized to be downstream effects of taxed working memory from rumination processes that impair reward learning. Furthermore, episodic memory impairments for positive information are hypothesized to be downstream effects of dopamine dysregulation from anhedonia in memory formation processes. The current study utilized the Diffusion Drift Model to examine whether anhedonia severity predicted differences in evidence accumulation using an index called drift rates in working and episodic memory tasks. One hundred and eight young adults completed a working memory task where they had to remember an abstract shape while presented with two positive, neutral, or negative distractors, followed by a surprise recognition test of the distractors the next day. Contrary to expectations, based on multivariate models, anhedonia severity did not predict differences in evidence accumulation rate. These results suggest that anhedonia symptoms may not be uniquely associated with memory processes for emotionally-valenced stimuli. Further studies should investigate the role of specific facets of anhedonia (e.g., anticipatory pleasure) and use different paradigms and neurophysiological measures to examine the proposed hypotheses
Hybrid Graph-Recurrent Architecture for Citation Recommendation via Future Embedding Forecasting
Get PDFThe rapid expansion of scientific literature has intensified the challenge of identifying relevant citations, particularly for newly published or under-cited papers. Traditional citation recommendation systems typically model static relationships or respond to past citation activity, offering limited predictive power for emerging works. In response, this thesis presents a temporal modeling framework for citation recommendation that anticipates future scholarly relevance by forecasting the latent representations of academic papers.
Building on prior work that utilized Temporal Graph Networks (TGNs) to model dynamic citation flows, we propose Graph-Time, a hybrid architecture that integrates a Graph Transformer with a GRU-based time series predictor. The transformer captures structural and temporal dependencies within yearly citation snapshots, while the GRU forecasts a paper’s future embedding based on its past trajectory. To address cold-start cases, We inc
We evaluate Graph-Time on the AMiner Citation Network V14, comprising over 5 million papers and 36 million citation links, using strict year-based splits to avoid information leakage. Compared to dynamic and static baselines, our model achieves substantial improvements in recommendation performance, including a 21 percent gain in MRR and notable gains in Precision@10 and Recall@10. Ablation studies confirm the contribution of each architectural component, and we discuss ethical considerations surrounding bias reinforcement in scholarly recommendations.
This work contributes a scalable, future-aware framework for citation forecasting and lays a foundation for real-time, temporally sensitive academic recommender systems
Defense Use of Digital Discovery in Criminal Cases: A Quantitative Analysis
No full textRecent criminal court reforms have required prosecutors to provide defense attorneys with broader and earlier discovery of evidence. For these discovery reforms to fulfill their aims of improved fairness and efficiency, defense attorneys must take advantage of the evidence disclosed by the prosecution. Prior studies suggest, however, that a range of factors, including low pay and high caseloads, impede effective defense representation in general. If similar factors hinder defense attorneys from reviewing discovery, discovery reforms would fail to meet their goals, and defendants would receive sub-standard representation. The recent adoption of digital evidence platforms by local jurisdictions allows us to study whether defense attorneys consistently fulfill their duty to review discovery. Analyzing data from digital evidence platforms used in felony cases in four Texas counties between 2018 and 2020, we examine whether and when defense attorneys fail to access evidence disclosed by the prosecution. We find that a substantial number of defense attorneys never access the discovery. The access rate varies by county, offense seriousness, attorney category, attorney experience, and file type. Drawing on review of prior scholarship and Bayesian analysis of the data, we discuss plausible interpretations of these variations