8383 research outputs found
Sort by
A Practical Application of Retrieval-Augmented Generation for Website-Based Chatbots: Combining Web Scraping, Vectorization, and Semantic Search
Get PDFThe Retrieval-Augmented Generation (RAG) model significantly enhances the capabilities of large language models (LLMs) by integrating information retrieval with text generation, which is particularly relevant for applications requiring context-aware responses based on dynamic data sources. This research study presents a practical implementation of a RAG model personalized for a Chabot that answers user inquiries from various specific websites. The methodology encompasses several key steps: web scraping using BeautifulSoup to extract relevant content, text processing to segment this content into manageable chunks, and vectorization to create embeddings for efficient semantic search. By employing a semantic search approach, the system retrieves the most relevant document segments based on user queries. The OpenAI API is then utilized to generate contextually appropriate responses from the retrieved information. Key results highlight the system's effectiveness in providing accurate and relevant answers, with evaluation metrics centered on response quality, retrieval efficiency, and user satisfaction. This research contributes a comprehensive integration of scraping, vectorization, and semantic search technologies into a cohesive chatbot application, offering valuable insights into the practical implementation of RAG models
Optimizing Stroke Prediction using Gated Recurrent Unit and Feature Selection in Sub-Saharan Africa
Get PDFBackground
Stroke remains a leading cause of death and disability worldwide, with African populations bearing a disproportionately high burden due to limited healthcare infrastructure. Early prediction and intervention are critical to reducing stroke outcomes. This study developed and evaluated a stroke prediction system using Gated Recurrent Units (GRU), a variant of Recurrent Neural Networks (RNN), leveraging the Afrocentric Stroke Investigative Research and Education Network (SIREN) dataset.
Method
The study utilized secondary data from the SIREN dataset, comprising 4,236 records with 29 phenotypes. Feature selection reduced these to 15 optimal phenotypes based on their significance to stroke occurrence. The GRU model, designed with 128 input neurons and four hidden layers (64, 32, 16, and 8 neurons), was trained and evaluated using 150 epochs, a batch size of 8, and metrics such as accuracy, AUC, and prediction time. Comparisons were made with traditional machine learning algorithms (Logistic Regression, SVM, KNN) and Long Short-Term Memory (LSTM) networks.
Results
The GRU-based system achieved a performance accuracy of 77.48%, an AUC of 0.84, and a prediction time of 0.43seconds, outperforming all other models. Logistic Regression achieved 73.58%, while LSTM reached 74.88% but with a longer prediction time of 2.23seconds. Feature selection significantly improved the model's performance compared to using all 29 phenotypes.
Conclusion
The GRU-based system demonstrated superior performance in stroke prediction, offering an efficient and scalable tool for healthcare. Future research should focus on integrating unstructured data, validating the model on diverse populations, and exploring hybrid architectures to enhance predictive accuracy
The home literacy environment in families with a history of reading difficulties
No full textIn this chapter, we consider the role of the home literacy environment (HLE) in the development of children at elevated risk of reading difficulties through a familial history of dyslexia. Characterized by difficulties in word-level literacy and underpinned by differences in phonological processing, dyslexia is both highly heritable and aetiologically complex. We first consider patterns of gene-environment interplay in reading development within a ‘multiple deficit’ framework. We review findings from prospective, longitudinal studies of dyslexia from various countries in relation to the HLE, before outlining findings from two studies in detail: The Wellcome Language and Reading Project in the UK, and the On Track study in Norway. There is growing evidence that a rich HLE can be protective in the reading trajectories of children at familial risk of dyslexia; and conversely, impoverished early literacy experiences combine with other risk factors to inhibit literacy trajectories. We consider the implications of this research base for the conceptualization of HLE and intervention design
Voices from Military Families: Young People's Reflections on Educational Experiences and Othering in British Schools
No full textThis paper critically examines the educational experiences of military children through the lens of Othering and exclusion, foregrounding young people's reflections alongside perspectives from educators and third-sector professionals. While inclusive education has gained policy traction, the distinct needs of military children remain underexplored in academic discourse. Their high mobility, exposure to deployment-related stress, and cultural distinctiveness often position them as outsiders within school communities. Drawing on in-depth interviews that centre on the lived experiences and voices of military young people, supplemented by insights from educators and support professionals, this study explores how everyday exclusion and systemic Othering manifest in classroom interactions, peer relationships, and institutional practices. Young people's reflections reveal frequent experiences of social and academic marginalisation, contributing to feelings of invisibility and disconnection from school environments. By privileging the narratives and understanding of military children, this research challenges prevailing assumptions about inclusion in schools. The analysis demonstrates how exclusion operates subtly within seemingly inclusive spaces, often overlooking the complex realities these young people navigate daily. The paper argues for targeted educational strategies that move beyond generic inclusion frameworks, advocating for approaches informed by the insights of military children to ensure their meaningful participation and a genuine sense of belonging in educational contexts
Artificial Intelligence in Computational and Materials Chemistry: Prospects and Limitations
Get PDFComputational chemistry, at the intersection of theoretical chemistry and computer science, employs various models to analyze molecular structures and properties, enabling the understanding and prediction of intricate chemical processes. The integration of artificial intelligence (AI) has revolutionized several fields, particularly in materials chemistry, with applications spanning drug discovery, materials design, and quantum mechanics. However, challenges related to quantum system complexity, model interpretability, and data quality remain a few of the Achilles’ heel of AI applications. This paper provides an overview of AI’s evolution in computational and materials chemistry, focusing on several applications. AI’s transformative potential in materials chemistry is emphasized, facilitating precise material property predictions, crucial for industries reliant on materials innovation. In materials chemistry, AI has led to substantial advancements, enabling the rapid discovery of materials with tailored properties. Yet, the challenges of modeling complex quantum systems, achieving model interpretability, and accessing high-quality data remain. The integration of AI into computational and materials chemistry promises to reshape the field, revolutionizing chemical research, materials design, and technological innovation. In order to harness AI’s full potential, transparent AI models, advanced quantum simulations, optimized data utilization, scalable computing, interdisciplinary collaboration, and ethical AI practices are essential
Trawling the LIPA Metaverse: Community Music, my LIPA, and Questions of Research and Scholarship
Get PDFExamining neuroanatomical correlates of win-stay, lose-shift behaviour
Get PDFThis study aimed to better understand the neuroanatomical correlates of decision-making strategies, particularly focusing on win-stay and lose-shift behaviours, using voxel-based morphometry (VBM) in a large cohort of healthy adults. Participants completed a forced-choice card-guessing task designed to elicit behavioural responses to rewards and losses. Using this task, we investigated the relationship between win-stay and lose-shift behaviour and both grey matter volume (GMV) and white matter volume (WMV). The frequency of win-stay and lose-shift behaviours was calculated for each participant and entered into VBM analyses alongside GMV and WMV measures. Our results revealed that increased lose-shift behaviour was associated with reduced GMV in key brain regions, comprising of the left superior temporal gyrus, right middle temporal gyrus, and the bilateral superior lateral occipital cortices. Interestingly, no significant associations were found between GMV or WMV, and win-stay behaviour. These results suggest that specific regions within the temporal and occipital lobes may be involved in modulating decision-making strategies following negative outcomes. Further analyses revealed that increased lose-shift behaviour was also associated with increased WMV in the left superior temporal gyrus. The absence of significant findings in relation to win-stay behaviour and the differential involvement of brain structures in lose-shift responses indicate that decision-making in the face of losses may involve distinct neuroanatomical mechanisms compared to decision-making following wins. This study advances our understanding of the structural brain correlates linked to decision-making strategies and highlights the complexity of brain-behaviour relationships in choice behaviour
Exploring link between skills, attitudes, and intentions in information technology industry: a study on entrepreneurial mindset
Get PDFThe study seeks to determine the relationship among entrepreneurial skills, attitude towards behaviour (ATB), and the entrepreneurial goals of IT professionals. With implications for comprehending entrepreneurship in a service-oriented business, this study explores how ATB connects abilities and intents using the theory of planned behaviour. With an emphasis on technology-driven niches, the study develops theoretical models and improves theory on how attitudes and skills interact to impact entrepreneurial intentions by incorporating ATB as a moderating variable of interest. Simple random sampling was used to pick the sample from a cross-sectional survey of 376 IT professionals, and partial least squares structural equation modelling (PLS-SEM) was used to investigate the hypothesized association between the variables. According to the results, ATB moderates the link between skills and intentions, suggesting that fostering both skill and intention development will support the growth of entrepreneurial ability. This realization has significant ramifications for politicians, educators, and practitioners in terms of establishing favourable environments, developing logical skill-building exercises, and encouraging optimistic entrepreneurial attitudes. According to identified factors, stakeholders have the ability to improve the climate for entrepreneurship and support economic growth. This study advances the theoretical understanding of entrepreneurship, particularly as it relates to the creation of intent in the technology sector
An Imbalance Regression Approach to Toxicity Prediction of Chemicals for Potential Use in Environmentally Acceptable Lubricants
Get PDFLubricants are complex mixtures of chemicals that help machines function at the right level of friction and wear. Lubricant formulation methods are based on empirical experience of chemical substances that have been used as lubricants for decades. In the last years, the discussion about their environmental problem has triggered new legislations resulting in the search for Environmentally Acceptable Lubricants, which should be biodegradable, minimally toxic, and nonbioaccumulative. Finding new chemicals that comply with these three criteria is a long and expensive process that can be boosted by machine learning (ML). In this paper, we are addressing toxicity prediction with machine learning models by exploring the application of ensemble learners to chemicals having imbalanced data distribution. We investigated the effectiveness of sampling techniques to balance the data and improve the performance of the ensemble learning model. The model can predict toxicity for nonundersampled groups, which in our case corresponds to the moderately to highly toxic groups. The results of this work are useful for lubricant formulators since regulations accept moderate-to-highly toxic chemicals in lubricants if their concentration is below 20 wt %