This review delves into using e-learning technology and personalized recommendation systems in education. It examines 60 articles from prominent databases and identifies the different methods used in recommendation systems, such as collaborative and content-based approaches with a recent shift towards machine learning. However, the current personalized recommendation system faces challenges such as a lack of understanding of the content, student discontinuity, language barriers, confusion in selecting study materials, inadequate infrastructure and funding. The review proposes using new digital technologies to address these issues, including Fluxy AI, Twin technology, AI-powered virtual proctoring, and Alter Ego. These technologies can create a dynamic and interactive learning environment, providing tailored learning experiences for students and insights for educators to provide targeted support and guidance. The integration of these technologies can improve individualized learning, increase understanding capacity and enhance the learning experience for students with speech disorders.
In recent years, the language of emotion has attracted widespread attention in music therapy. Each piece of music inherently carries an emotional essence, making the study of music emotion crucial. This work begins by exploring the historical background of music emotion and then conducts a comprehensive survey of existing machine learning, deep learning, and ensemble techniques employed in music emotion detection, along with a detailed analysis of the datasets utilized in these methodologies. Additionally, we identify and discuss the key issues and challenges emerging from our survey. Our work also delves into the contemporary technology of music emotion detection, focusing specifically on electroencephalography (EEG) signals. Lastly, we outline potential avenues for future research in music emotion detection.
Depending on the severity and spatial-temporal variability, droughts can have a wide range of impacts such as crop failure, water shortages, and food insecurity. Accurate and timely forecasting is necessary to mitigate the hazards of extreme weather events, such as droughts, brought on by climate change. A district like Chitradurga in India, which typically receives around 450-600mm of annual rainfall, will require advanced drought mitigation strategies and plans before the onset of the drought. This research focuses on 1-step lead time forecasting of meteorological drought episodes making use of the 6-month Standardised Precipitation Index (SPI-6) as indicator. The fine resolution rainfall data (0.25° × 0.25°) obtained from the Indian Meteorological Department was used to derive the 6-month SPI data of 23 grid stations. The 1-step lead time SPI-6 time series was forecast considering the antecedent SPI-6 time series data as model input. The Mutual Information was used to determine the most relevant input features for drought forecasting. The standard Artificial Neural Network, an advanced machine learning framework - Multivariate Adaptive Regression Splines, and the ensemble learning-based CatBoost Regression and Gradient Tree Boosting paradigms were employed to forecast drought episodes. Error and efficiency metrics were employed for performance evaluation of the simulated models. The multivariate adaptive regression splines and gradient tree boosting forecasts had slightly higher accuracy and lower error rates than the artificial neural network model, which suggests that they may be more reliable for drought forecasting. The root mean square error and normalized Nash-Sutcliffe efficiency ranges of the multivariate adaptive regression splines model (during test phase) were 0.37–0.54 and 0.78–0.87, respectively. The thematic maps that were created using spatial interpolation of model forecasts from all the stations also confirmed that the district as a whole experienced drought in April 2019.
24 February 2024 in International Journal of Information Technology
In this work we introduce Split-Fed Learning (SFL), a new framework that combines the concepts of Federated Learning (FL) and Split Learning (SL), to provide privacy-aware and scalable training of machine learning models in settings with distributed data. With organizations needing more and more to take advantage of decentralized data sources without compromising on security, the potential of using SFL becomes all the more apparent. We performed extensive experiments to demonstrate the superiority of SFL against FL (both traditional and naive ones) as well as SL with various datasets. Experimental results show that SFL can reduce the communication overhead and training time by a large margin while achieving a competitive accuracy of final model. In contrast, SFL delivered 30% better training speed and 15% accuracy than baseline models particularly for non-IID data distributions challenges. Lastly, a thorough breakdown of the SFL architecture explains its novelty for separating model training between clients and servers with respect to model partitioning. Hence, this optimization not only improves the efficiency of resource utilization but also boosts the resiliency of learning process against data heterogeneity. These results demonstrate the potential of SFL as a scalable and efficient approach to privacy-preserving machine learning, providing an alternative solution to recent challenges faced by organizations that deal with sensitive data. This is the first paper establishing such type of literature for SFL, representing an initial step toward a theoretical background for future research for implementing distributed learning frameworks and next towards practical usage scenarios on testing SFL.
The rapid spread of false information on social media has become a major challenge in today's digital world. This has created a need for an effective rumor detection system that can identify and control the spread of false information in real-time. The proposed work introduces a rumor detection system by integrating transformer-based models such as BERT, DistilBERT, and TinyBERT with traditional Machine Learning (ML) techniques. The proposed method involves fine-tuning transformer models to recognize the contextual meaning of social media posts. The classifiers include Decision Tree (DT), Support Vector Machine (SVM), Random Forest (RF) and Naïve Bayes (NB), which help in categorizing content as either rumor or non-rumor based on the patterns. The proposed work evaluated BERT, DistilBERT, TinyBERT combined with ML models (SVM, DT, RF, NB) across PHEME dataset using 70:30, 60:40, and 80:20 splits. Overall, BERT+DT and TinyBERT+SVM provided significant results, with BERT+RF and DistilBERT+NB demonstrating better classification capabilities across various events and split ratios on the dataset.