Hallucination Reduction in Large Language Models with Retrieval-Augmented Generation Using Wikipedia Knowledge

Опубликована: Май 30, 2024

Natural language understanding and generation have seen great progress, yet the persistent issue of hallucination undermines reliability model outputs. Introducing retrieval-augmented (RAG) with external knowledge sources, such as Wikipedia, presents a novel significant approach to enhancing factual accuracy coherence in generated content. By dynamically integrating relevant information, Mistral demonstrates substantial improvements precision, recall, overall quality responses. This research offers robust framework for mitigating hallucinations, providing valuable insights deploying reliable AI systems critical applications. The comprehensive evaluation underscores potential RAG advance performance trustworthiness large models.

Язык: Английский

---

Bring Retrieval Augmented Generation to Google Gemini via External API: An Evaluation with BIG-Bench Dataset

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Май 10, 2024

The integration of Retrieval Augmented Generation (RAG) into existing large language models represents a significant shift towards more dynamic and context-aware AI systems. In this work, Google Gemini, state-of-the-art model, has been enhanced with RAG capabilities to leverage external, real-time data sources during the response generation process. This augmentation aims address traditional limitations models, particularly in generating responses that require up-to-date information adaptability complex user queries. performance RAG-enhanced Gemini was rigorously evaluated using BIG-Bench dataset, which includes tasks designed test bounds terms reasoning, contextuality, factual accuracy. Quantitative results from evaluation demonstrate marked improvements accuracy contextual relevance across various tasks, indicating effectiveness enhancing model performance. Qualitative assessments further support these findings, highlighting model’s improved ability generate precise relevant responses. However, also introduces challenges related computational efficiency scalability, emphasizing need for optimization. paper discusses potential future research directions, including application other datasets, exploration different configurations, development sophisticated handling techniques enhance applicability. ongoing advancement technologies promises significantly broaden utility AI-driven systems real-world applications, making them adaptable useful diverse scenarios.

Язык: Английский

---

Measuring the Perceived IQ of Multimodal Large Language Models Using Standardized IQ Tests

Опубликована: Май 13, 2024

Evaluating the intelligence of multimodal large language models (LLMs) using adapted human IQ tests poses unique challenges and opportunities for understanding AI capabilities.By applying Wechsler Adult Intelligence Scale (WAIS), customized to assess cognitive functions LLMs such as Baidu Benie, Google Gemini, Anthropic Claude, significant insights into complex intellectual landscape these systems were revealed.The study demonstrates that can exhibit sophisticated abilities, performing tasks requiring advanced verbal comprehension, perceptual reasoning, problemsolving-traditionally considered within purview cognition.The research also highlights distinct profiles each model, reflecting their specialized architectures training.However, acknowledges inherent limitations in human-oriented assessment, emphasizing need ongoing refinement testing methodologies keep pace with development.Future directions include creation dynamic adaptive frameworks better align capabilities evolving systems, ensuring integration societal remains aligned values safety standards.

Язык: Английский

---

The Effect of the Use of Interactive Media in Arabic Language Learning on Students' Learning Outcomes at Nurul Ilmi Integrated Islamic Elementary School, Jambi

Asian Journal of Education and Social Studies, Год журнала: 2025, Номер 51(1), С. 158 - 164

Опубликована: Янв. 11, 2025

Interactive media is one of the essential elements in improving quality education. This study aims to determine effect use interactive Arabic language learning on outcomes class 2A students at Nurul Ilmi Integrated Islamic Elementary School Jambi. The approach used this quantitative with a quasi-experimental design. population was entire second grade, totalling 236 people, divided into 10 group classes. sample 20 students, who were taken randomly. data collected using test techniques (pretest and post-test) from perspective learning. analysis N-gain T-test. results showed that positively significantly affected outcomes. increase average "high" category. T-test significant. Thus, it recommended multimedia-based teaching materials be developed support implementation more professional

Язык: Английский

---

Metodología para generación de entorno de realidad virtual de espacio turístico para alcanzar competencias de estudiantes de Turismo y Hotelería

European Public & Social Innovation Review, Год журнала: 2025, Номер 10, С. 1 - 23

Опубликована: Янв. 21, 2025

Introducción: La realidad virtual (RV) es una herramienta útil en la formación universitaria, especialmente carreras como Turismo y Hotelería (TH). Este estudio propone metodología para crear un entorno de RV espacio turístico, permitiendo a los estudiantes familiarizarse con él sin necesidad viajes presenciales. Metodología: Se desarrolla sistema adaptado al modelo pedagógico conectivista, que facilita el aprendizaje TH. integra turístico delimitado, interactuar repetidamente hasta alcanzar objetivos aprendizaje. Resultados: creación entornos permite espacios turísticos, superar barreras recursos acceder lugares condiciones preservación. Además, evaluar su medir mejora sus conocimientos. Discusión: propuesta demuestra ser efectiva simular visitas estudio, optimizando proceso educativo repetir experiencia interiorizar contenidos. Conclusiones: integración ofrece alternativa valiosa limitaciones logísticas mejorar práctico.

---

A systematic literature review to implement large language model in higher education: issues and solutions

Discover Education, Год журнала: 2025, Номер 4(1)

Опубликована: Фев. 15, 2025

Язык: Английский

---

Can AI teach me employability? A multi-national study in three countries

Frontiers in Artificial Intelligence, Год журнала: 2024, Номер 7

Опубликована: Ноя. 18, 2024

This paper examines the impact of using an Artificial Intelligence (AI) teacher for current Higher Education (HE) students from three countries. The study utilized AI avatar powered by a fine-tuned Large Language Model (LLM), OIMISA, which is trained solely teaching and learning applications. provided 9-lesson course on employability transferable skills. In total 207 across institutions enrolled in programme. results demonstrate noteworthy completion rate over 47%, along with high levels engagement all student cohorts satisfaction rates students. These show potential AI-based virtual teachers countries HE compared to use MOOC platforms.

Язык: Английский

---

Orbit Weighting Scheme in the Context of Vector Space Information Retrieval

Computers, materials & continua/Computers, materials & continua (Print), Год журнала: 2024, Номер 80(1), С. 1347 - 1379

Опубликована: Янв. 1, 2024

This study introduces the Orbit Weighting Scheme (OWS), a novel approach aimed at enhancing precision and efficiency of Vector Space information retrieval (IR) models, which have traditionally relied on weighting schemes like tf-idf BM25. These conventional methods often struggle with accurately capturing document relevance, leading to inefficiencies in both performance index size management. OWS proposes dynamic mechanism that evaluates significance terms based their orbital position within vector space, emphasizing term relationships distribution patterns overlooked by existing models. Our research focuses evaluating OWS's impact model accuracy using Information Retrieval metrics Recall, Precision, Interpolated Average Precision (IAP), Mean (MAP). Additionally, we assess effectiveness reducing inverted size, crucial for efficiency. We compare OWS-based models against others different schemes, including variations BM25Delta. Results reveal superiority, achieving 54% Recall 81% MAP, notable 38% reduction size. highlights potential optimizing processes underscores need further this underrepresented area fully leverage capabilities methodologies.

Язык: Английский

---

"Let’s ask what AI thinks then!": Using LLMs for Collaborative Problem-Solving in Virtual Environments

2022 IEEE International Symposium on Mixed and Augmented Reality Adjunct (ISMAR-Adjunct), Год журнала: 2024, Номер unknown, С. 193 - 198

Опубликована: Окт. 21, 2024

Язык: Английский

---