The rapid proliferation of deepfake videos, generated using advanced machine learning techniques to create highly realistic but misleading content, poses significant challenges across various sectors, including cybers...
详细信息
The rapid proliferation of deepfake videos, generated using advanced machine learning techniques to create highly realistic but misleading content, poses significant challenges across various sectors, including cybersecurity, media integrity, and personal privacy. Detecting these deepfakes has become essential for maintaining trust in digital media and preventing malicious exploitation. This paper presents a novel approach to deepfake video detection by employing the AdaBoost algorithm, a powerful ensemble learning method recognized for its ability to improve classification performance by focusing on difficult-to-classify instances. Using the Deepfake Detection Challenge (DFDC) dataset, our study demonstrates that the AdaBoost classifier, when coupled with a carefully designed feature set, achieves competitive accuracy in detecting deepfake videos. Our results show that this approach provides an effective solution for deepfake detection, with strong recall performance, making it a viable method for real-world applications.
Audio-to-talking face generation stands at the forefront of advancements in generative AI. It bridges the gap between audio and visual representations by generating synchronized and realistic talking faces. Despite re...
详细信息
Early detection and diagnosis are crucial because cardiovascular diseases are the leading cause of fatalities in the world. Timely diagnosis and intervention can significantly improve patient outcomes and reduce healt...
详细信息
Early detection and diagnosis are crucial because cardiovascular diseases are the leading cause of fatalities in the world. Timely diagnosis and intervention can significantly improve patient outcomes and reduce healthcare costs. In this study, we used Logistic Regression, Random Forest, Decision Tree, K-Nearest Neighbour, and Support Vector Machines to predict the risk of cardiovascular diseases based on two datasets obtained from the UCI library. We assessed each algorithm’s performance based on its accuracy, precision, recall, and F1 Score. Our findings show that KNN performed better than the other algorithms, with Accuracy of 90.16%. This study shows how machine learning algorithms can be used to predict cardiac diseases and offers guidance for further study in this area.
Variational Bayesian learning (VBL)-aided extended target localization is conceived for orthogonal frequency division multiplexing (OFDM) based-mmWave MIMO systems using the OFDM integrated sensing and communication (...
详细信息
The quality of the air is crucial for people who suffer from respiratory conditions. The prediction of the air quality index (AQI) is essential for preventing health issues, particularly for those with respiratory dis...
详细信息
The quality of the air is crucial for people who suffer from respiratory conditions. The prediction of the air quality index (AQI) is essential for preventing health issues, particularly for those with respiratory diseases. Deteriorating air quality is often linked to increased residential, industrial, and transportation activity. Accurate AQI prediction can help in mitigating these health risks. Disseminating information about the air quality index among diverse populations is crucial. The motive of this study is to use several machine learning regression and classification approaches to develop quantitative and accurate air quality forecast models. Incorporating a preventive system that provides preventive measures to healthy people as well as those suffering from cancer and respiratory diseases is an additional component of this. This study evaluates multiple supervised machine learning algorithms, including Random Forest and XGBoost, for AQI prediction using real-time data from four Indian cities. To identify the most effective model, five different machine learning models were used, and hyperparameter tuning was performed to achieve better accuracy. After evaluating the R2 score, root mean squared error (RMSE), mean absolute error (MAE), and mean squared error (MSE) for five different regression algorithms, as well as precision, accuracy, recall, and F1 score for classification algorithms, it was found that the Random Forest and XGBoost (using the best hyperparameters) classifiers performed best among other models. The Random Forest model achieved an MAE of 0.97 and an R-squared of 0.998, while the XGBoost classifier reached a perfect accuracy of 1.0 after hyperparameter tuning, providing valuable insights for respiratory care and preventive measures. Other models that were looked at included k-nearest neighbors (KNN), decision trees, and support vector machines (SVM). When visiting locations with high air quality index (AQI) values, exercise carefulness since
Mobile technology use in education is changing, much like in the business and health sectors. Research is now conducted on designing user-centric platforms that allow people to engage in teaching and learning activiti...
详细信息
In recent days, deep learning technologies have gained more and more interest in computer related task. In generative models, autoencoder (AE) has achieved a tremendous success in many fields, especially in image gene...
详细信息
Mental health issues, including anxiety, stress, and depression, may remain untreated until they escalate to a severe level. The issues significantly impact an individual's overall well-being and productivity. Tim...
详细信息
Tiny machine learning (TinyML) is an essential component of emerging smart microcontrollers (MCUs). However, the protection of the intellectual property (IP) of the model is an increasing concern due to the lack of de...
ISBN:
(纸本)9798350323481
Tiny machine learning (TinyML) is an essential component of emerging smart microcontrollers (MCUs). However, the protection of the intellectual property (IP) of the model is an increasing concern due to the lack of desktop/server-grade resources on these power-constrained devices. In this paper, we propose STML, a system and algorithm co-design to Secure IP of TinyML on MCUs with ARM TrustZone. Our design jointly optimizes memory utilization and latency while ensuring the security and accuracy of emerging models. We implemented a prototype and benchmarked with 7 models, demonstrating STML reduces 40% of model protection runtime overhead on average.
Breadth-first search (BFS) is a fundamental graph algorithm that presents significant challenges for parallel implementation due to irregular memory access patterns, load imbalance and synchronization overhead. In thi...
详细信息
暂无评论