A comprehensive review of large language models: issues and solutions in learning environments

Discover Sustainability, Год журнала: 2025, Номер 6(1)

Опубликована: Янв. 14, 2025

A significant advancement in artificial intelligence is the development of large language models (LLMs). Despite opposition and explicit bans by some authorities, LLMs continue to play a transformative role, particularly education, improving understanding generation capabilities. This study explores LLMs' types, history, training processes, alongside their application including digital higher education settings. novel theoretical framework proposed guide integration into addressing key challenges such as personalization, ethical concerns, adaptability. Furthermore, presents practical case studies solutions barriers, data privacy bias, offering insights role enhancing teaching–learning process. By providing systematic analysis proposing structured framework, this advances current knowledge highlights potential revolutionizing education.

Язык: Английский

---

Analyzing Sentiments Regarding ChatGPT Using Novel BERT: A Machine Learning Approach

Information, Год журнала: 2023, Номер 14(9), С. 474 - 474

Опубликована: Авг. 25, 2023

Chatbots are AI-powered programs designed to replicate human conversation. They capable of performing a wide range tasks, including answering questions, offering directions, controlling smart home thermostats, and playing music, among other functions. ChatGPT is popular AI-based chatbot that generates meaningful responses queries, aiding people in learning. While some individuals support ChatGPT, others view it as disruptive tool the field education. Discussions about this can be found across different social media platforms. Analyzing sentiment such data, which comprises people’s opinions, crucial for assessing public regarding success shortcomings tools. This study performs analysis topic modeling on ChatGPT-based tweets. tweets author’s extracted from Twitter using hashtags, where users share their reviews opinions providing reference thoughts expressed by The Latent Dirichlet Allocation (LDA) approach employed identify most frequently discussed topics relation For analysis, deep transformer-based Bidirectional Encoder Representations Transformers (BERT) model with three dense layers neural networks proposed. Additionally, machine learning models fine-tuned parameters utilized comparative analysis. Experimental results demonstrate superior performance proposed BERT model, achieving an accuracy 96.49%.

Язык: Английский

---

Global insights and the impact of generative AI-ChatGPT on multidisciplinary: a systematic review and bibliometric analysis

Connection Science, Год журнала: 2024, Номер 36(1)

Опубликована: Май 16, 2024

In 2022, OpenAI's unveiling of generative AI Large Language Models (LLMs)- ChatGPT, heralded a significant leap forward in human-machine interaction through cutting-edge technologies. With its surging popularity, scholars across various fields have begun to delve into the myriad applications ChatGPT. While existing literature reviews on LLMs like ChatGPT are available, there is notable absence systematic (SLRs) and bibliometric analyses assessing research's multidisciplinary geographical breadth. This study aims bridge this gap by synthesising evaluating how has been integrated diverse research areas, focussing scope distribution studies. Through review scholarly articles, we chart global utilisation scientific domains, exploring contribution advancing paradigms adoption trends among different disciplines. Our findings reveal widespread endorsement multiple fields, with implementations healthcare (38.6%), computer science/IT (18.6%), education/research (17.3%). Moreover, our demographic analysis underscores ChatGPT's reach accessibility, indicating participation from 80 unique countries ChatGPT-related research, most frequent keyword occurrence, USA (719), China (181), India (157) leading contributions. Additionally, highlights roles institutions such as King Saud University, All Institute Medical Sciences, Taipei University pioneering dataset. not only sheds light vast opportunities challenges posed pursuits but also acts pivotal resource for future inquiries. It emphasises that (LLM) role revolutionising every field. The insights provided paper particularly valuable academics, researchers, practitioners disciplines, well policymakers looking grasp extensive impact technologies community.

Язык: Английский

---

Overviewing Biases in Generative AI-Powered Models in the Arabic Language

Advances in computational intelligence and robotics book series, Год журнала: 2025, Номер unknown, С. 361 - 390

Опубликована: Фев. 28, 2025

Natural Language Processing (NLP) is an emerging field often integrated into Artificial Intelligence (AI) technologies. NLP has significantly advanced, leading to the widespread use of generative AI-powered (Gen-AI) models across various domains. However, while Gen-AI systems have been successfully implemented in several languages, AI-based language still face considerable challenges and shortcomings, including generating biases sensitive languages like Arabic. Therefore, primary objective this chapter provide overview Gen-AI-powered context Arabic language, exploring sources these biases, their implications, potential strategies for mitigation. The underscore need ongoing research development create more equitable accurate AI systems. By understanding origins implications implementing effective mitigation strategies, we can work towards that better serve diverse linguistic communities.

Язык: Английский

---

Multilingual hope speech detection: A Robust framework using transfer learning of fine-tuning RoBERTa model

Journal of King Saud University - Computer and Information Sciences, Год журнала: 2023, Номер 35(8), С. 101736 - 101736

Опубликована: Авг. 29, 2023

Hope Speech Detection (HSD) from social media is a new direction for promoting and supporting positive content to encourage harmony positivity in society. As users of belong different linguistic communities, hope speech detection rarely studied as multilingual task considering low-resource languages. Moreover, prior studies explored only monolingual techniques, the Russian language not addressed. This study tackles issue Multi-lingual (MHSD) English languages using transfer learning paradigm with fine-tuning approach. We explore joint multi-lingual translation-based approaches tackle multilingualism, where latter approach adopts translation mechanism transform all into one then classify them. The method handles it by designing universal classifier various strengths Robustly Optimized BERT Pre-Training Approach (RoBERTa) that showed benchmark capturing semantics contextual information within content. proposed framework consists several stages: 1) data preprocessing, 2) representation RoBERTa models, 3) phase, 4) classification two labels. A corpus built, containing YouTube comments. Several experiments are conducted semi-supervised bilingual datasets. findings show demonstrated performance outperformed baselines. Furthermore, (Russian-RoBERTa) offered best achieving 94% accuracy 80.24% f1-score.

Язык: Английский

---

Research on industry label recognition method based on RoBERTa driven by data

International Journal of Low-Carbon Technologies, Год журнала: 2025, Номер 20, С. 626 - 634

Опубликована: Янв. 1, 2025

Abstract This study introduces an enhanced RoBERTa-based model, called Industry Aware RoBERTa (IA-RoBERTa), designed to improve the accuracy and generalization of industry label recognition. IA-RoBERTa innovatively integrates structured knowledge through a graph fusion approach, using multigranularity input representation industry-aware self-attention mechanisms. Together, these features enhance model’s ability efficiently process understand industry-specific information. In addition, includes layered classifier that expertly handles fine-grained categories. Experimental evaluations recognition datasets show outperforms existing methods in terms accuracy, F1 scores, macro-average performance metrics.

Язык: Английский

---

Fake news detection: state-of-the-art review and advances with attention to Arabic language aspects

PeerJ Computer Science, Год журнала: 2025, Номер 11, С. e2693 - e2693

Опубликована: Март 11, 2025

The proliferation of fake news has become a significant threat, influencing individuals, institutions, and societies at large. This issue been exacerbated by the pervasive integration social media into daily life, directly shaping opinions, trends, even economies nations. Social platforms have struggled to mitigate effects news, relying primarily on traditional methods based human expertise knowledge. Consequently, machine learning (ML) deep (DL) techniques now play critical role in distinguishing necessitating their extensive deployment counter rapid spread misinformation across all languages, particularly Arabic. Detecting Arabic presents unique challenges, including complex grammar, diverse dialects, scarcity annotated datasets, along with lack research field detection compared English. study provides comprehensive review examining its types, domains, characteristics, life cycle, approaches. It further explores recent advancements leveraging ML, DL, transformer-based for detection, special attention delves Arabic-specific pre-processing techniques, methodologies tailored language, datasets employed these studies. Additionally, it outlines future directions aimed developing more effective robust strategies address challenge content.

Язык: Английский

---

The Intersection of Generative AI and Healthcare: Addressing Challenges to Enhance Patient Care

Опубликована: Март 3, 2024

Язык: Английский

---

An Automatic Sentiment Analysis Method for Short Texts Based on Transformer-BERT Hybrid Model

IEEE Access, Год журнала: 2024, Номер 12, С. 93305 - 93317

Опубликована: Янв. 1, 2024

Sentiment analysis towards short texts is always facing challenges, because only contain limited semantic characteristics. As a result, this paper constructs specific large language structure to deal with issue. In all, novel automatic sentiment method for based on Transformer-BERT hybrid model proposed by paper. Firstly, BERT utilized extract word vectors, and integrated topic vectors improve textual feature expression ability. Then, the fused are input into Bidirectional Gated Recurrent Unit (Bi-GRU) learn contextual features. part, Transformer applied behind Bi-GRU combined previous module output results. addition, Accuracy, Precision, Recall F1 indexes were collected from real-world Twitter datasets shopping data evaluate performance of method. The experimental results show that performs well in many indexes. Compared traditional method, it has achieved remarkable improvement, achieves higher accuracy efficiency texts, good generalization

Язык: Английский

---

Modeling group-level public sentiment in social networks through topic and role enhancement

Knowledge-Based Systems, Год журнала: 2024, Номер 305, С. 112594 - 112594

Опубликована: Окт. 11, 2024

Язык: Английский

---