Diabetic retinopathy, often resulting from conditions like diabetes and hypertension, is a leading cause of blindness globally. With diabetes affecting millions worldwide and anticipated to rise significantly, early detection becomes paramount. The survey scrutinizes existing literature, revealing a noticeable absence of consideration for computational complexity aspects in deep learning models. Notably, most researchers concentrate on employing deep learning models, and there is a lack of comprehensive surveys on the role of vision transformers in enhancing the efficiency of these models for DR detection. This study stands out by presenting a systematic review, exclusively considering 84 papers published in reputable academic journals to ensure a focus on mature research. The distinctive feature of this Systematic Literature Review (SLR) lies in its thorough investigation of computationally efficient approaches and models for DR detection. It sheds light on the incorporation of vision transformers into deep learning models, highlighting their significant contribution to improving accuracy. Moreover, the research outlines clear objectives related to the identified problem, giving rise to specific research questions. Following an assessment of relevant literature, data is extracted from digital archives. Additionally, in light of the results obtained from this SLR, a taxonomy for the detection of diabetic retinopathy has been presented. The study also highlights key research challenges and proposes potential avenues for further investigation in the field of detecting diabetic retinopathy.

In today’s digital era, the escalating phenomenon of cyberbullying is a pervasive and growing concern. With the increasing prevalence of social media platforms, such as Twitter, online abusive behavior has become a significant issue that often leads to unpleasant experiences for users. Manual detection of abnormal and bullying behavior within the realm of social media is inherently not scalable. Moreover, most existing studies on cyberbullying detection have been predominantly conducted in English and very limited work has been done on Urdu (a widely used language in Asia). This paper presents an approach for detecting cyberbullying in Roman Urdu tweets and identifying abuser profiles on Twitter. Firstly, we develop a text corpus of Roman Urdu tweets with user profile data. Subsequently, we employ Gated Recurrent Unit (GRU) model coupled with the application of word2vec technique for word embedding to develop a cyberbullying detection model. Furthermore, we present temporal abusive tweet probability analysis method to provide a nuanced analysis of the number of bullying and non-bullying tweets sent by individuals within a specific time interval. To evaluate the performance, we compare the GRU-based approach with other machine learning models. The results show that the GRU model with lexical normalization gives the best results with an accuracy of 97% and F1-measure of 97%.

Topic modeling enables the discovery of concealed themes and patterns in extensive text collections. It facilitates a thorough examination of the messages present in religious texts. Topic modeling for Quranic verses is a trending study area, with various translations already explored including Bahasa, English, and Arabic. Yet, there is a need for further research, particularly in Urdu translations of the Quran. In this study, we propose applying the BERTopic framework to Urdu translations of the Holy Quran. By leveraging the BERTopic approach, which incorporates a fine-tuned BERT model, we aim to capture the contextual nuances and linguistic complexities unique to the Quran. In this study, we utilized existing Urdu translations of the Quran from eight different translators sourced from Tanzil, a renowned resource for Quranic text and translations. We assessed the performance of our proposed BERTopic model compared to traditional techniques like LDA and NMF, using coherence and diversity metrics. The results indicate that our BERT-based approach outperforms these conventional methods, achieving an average coherence improvement of 0.03 and a diversity score of 0.83. These findings highlight the effectiveness of BERTopic in extracting meaningful topics from Urdu translations of the Holy Quran and contribute to the computational analysis of religious texts, supporting scholarly endeavors in comparative studies of Quranic translations in Urdu.

The exponential growth of digital content has made it increasingly difficult for users to understand the information, particularly in domains like political news. The study provides a framework for domain-specific multi-document political news summarization by using the transformer-based models that are fine-tuned on the synthetic dataset. The use of a generative Artificial Intelligence (AI) model, specifically the Bidirectional and Autoregressive Transformer (BART) model, for summarizing multi-document political news was explored in this paper. To improve the summarization process, real-world data was augmented, and a synthetic dataset was generated using the Adversarial Contextualized Text - Generative Adversarial Network (ACT-GAN) model to create a diverse and domain-specific training dataset. All four models, including BART, Pegasus, Text-To-Text Transfer Transformer (T5), and Seq2Seq, were fine-tuned and evaluated using a synthetic dataset generated via ACT-GAN and benchmarked using ROUGE metrics. BART performs the best in accuracy, recall, and F1-score, showing it can create clear and logical summaries while keeping the important context and structure intact. The results show that the transformer with the synthetic data helps to improve the quality of the political data summarization. In addition to addressing the challenges of extensive information, this framework also employs generative AI models in text summarization domains.

Machine Learning (ML) significantly drives the advancement of smart cities. This survey, using databases like IEEE Explorer, Web of Sciences, and Google Scholar, thoroughly investigated 22 papers published between 2021 and 2023. The focus was on identifying the prevalent ML models in smart cities and the specific sub-areas capturing the most attention. The study says that out of 22 research papers, about 63% used supervised learning techniques for smart city applications. The most common models were Naive Bayes and Support Vector Machines, especially in the areas of transportation, energy, environment, and healthcare. The industry has significance, due to its potential for conversion, especially with the urbanization of rural areas. This highlights the necessity for extensive future advancements. The results of this survey about the significance of machine learning in smart cities give us a path that will demand ongoing innovation to ensure the sustainable growth of both urban and rural areas going forward. Using machine learning, we can not only enhance the productivity of the city system but also increase the efficiency in diverse aspects of urban life.