A Review on Large Language Models: Architectures, Applications, Taxonomies, Open Issues and Challenges

Abstract

<p>Large Language Models (LLMs) recently demonstrated extraordinary capability, including natural language processing (NLP), language translation, text generation, question answering, etc. Moreover, LLMs are a new and essential part of computerized language processing, having the ability to understand complex verbal patterns and generate coherent and appropriate replies for the situation. Though this success of LLMs has prompted a substantial increase in research contributions, rapid growth has made it difficult to understand the overall impact of these improvements. Since a lot of new research on LLMs is coming out quickly, it is getting tough to get an overview of all of them in a short note. Consequently, the research community would benefit from a short but thorough review of the recent changes in this area. This article thoroughly overviews LLMs, including their history, architectures, transformers, resources, training methods, applications, impacts, challenges, etc. This paper begins by discussing the fundamental concepts of LLMs with its traditional pipeline of the LLM training phase. It then provides an overview of the existing works, the history of LLMs, their evolution over time, the architecture of transformers in LLMs, the different resources of LLMs, and the different training methods that have been used to train them. It also demonstrated the datasets utilized in the studies. After that, the paper discusses the wide range of applications of LLMs, including biomedical and healthcare, education, social, business, and agriculture. It also illustrates how LLMs create an impact on society and shape the future of AI and how they can be used to solve real-world problems. Then it also explores open issues and challenges to deploying LLMs in real-world aspects, including ethical issues, model biases, computing resources, interoperability, contextual constraints, privacy, security, etc. It also discusses methods to improve the robustness and controllability of LLMs. Finally, the study analyses the future of LLM research and issues that need to be overcome to make LLMs more impactful and reliable. However, this review paper aims to help practitioners, researchers, and experts thoroughly understand the evolution of LLMs, pre-trained architectures, applications, challenges, and future goals. Furthermore, it serves as a valuable reference for future development and application of LLM in numerous practical domains.</p>

Comparative Analysis of ChatGPT-4 and Google Gemini for Spam Detection on the SpamAssassin Public Mail Corpus

“…Comparative studies between AI models, specifically Chat-GPT and Google AI tools, have been a focal point for understanding the strengths and limitations of different approaches in spam detection. Investigations into the accuracy of these models in text classification tasks have shown varied results, with some studies indicating the superior performance of Google's AI in understanding context, while others highlight ChatGPT's ability to generate and analyze text [51], [52], [53]. Research on the use of these AI tools in detecting phishing emails has presented mixed findings, with some models excelling in identifying sophisticated phishing techniques [25], [54].…”

“…Research on the use of these AI tools in detecting phishing emails has presented mixed findings, with some models excelling in identifying sophisticated phishing techniques [25], [54]. The efficiency of ChatGPT and Google AI models in processing and classifying large volumes of data has been compared, noting differences in speed and computational resource requirements [51], [55]. Studies focusing on the adaptability of these models to new forms of spam have underscored the importance of continuous learning capabilities [56], [57].…”

“…Studies focusing on the adaptability of these models to new forms of spam have underscored the importance of continuous learning capabilities [56], [57]. The effectiveness of pre-trained models versus models trained specifically for spam detection tasks has been debated, with insights into the benefits and drawbacks of each approach [56], [51], [35], [58]. The role of user feedback in improving model accuracy has been examined, showing how interactive learning can enhance performance [51], [56], [59].…”

“…The effectiveness of pre-trained models versus models trained specifically for spam detection tasks has been debated, with insights into the benefits and drawbacks of each approach [56], [51], [35], [58]. The role of user feedback in improving model accuracy has been examined, showing how interactive learning can enhance performance [51], [56], [59]. Lastly, the impact of data privacy concerns on the deployment of these AI tools for spam detection has been a critical area of discussion, reflecting on the balance between effectiveness and user trust [59], [60], [61], [62].…”