Search

Recently, the number of Internet of Things (IoT)-connected devices has increased daily. Consequently, cybersecurity challenges have increased due to the natural diversity of the IoT, limited hardware resources, and limited security capabilities. Intrusion detection systems (IDSs) play a substantial role in securing IoT networks. Several researchers have focused on machine learning (ML) and deep learning (DL) to develop intrusion detection techniques. Although ML is good for classification, other methods perform better in feature transformation. However, at the level of accuracy, both learning techniques have their own certain compromises. Although IDSs based on ML and DL methods can achieve a high detection rate, the performance depends on the training dataset size. Incidentally, collecting a large amount of data is one of the main drawbacks that limits performance when training datasets are lacking, and such methods can fail to detect novel attacks. Few-shot learning (FSL) is an emerging approach that is employed in different domains because of its proven ability to learn from a few training samples. Although numerous studies have addressed the issues of IDSs and improved IDS performance, the literature on FSL-based IDSs is scarce. Therefore, an investigation is required to explore the performance of FSL in IoT IDSs. This work proposes an IoT intrusion detection model based on a convolutional neural network as a feature extractor and a prototypical network as an FSL classifier. The empirical results were analyzed and compared with those of recent intrusion detection approaches. The accuracy results reached 99.44%, which shows a promising direction for involving FSL in IoT IDSs. [ABSTRACT FROM AUTHOR]

The cultivation of creativity is closely related to language learning. How to design the creativity promotion mechanism of English teaching in the public environment is the challenge faced by English teachers. With the advent of the era of big data, English teachers can apply the latest research results to classroom teaching. For example, the rational use of social media helps students to learn and communicate in the language, cultivate students' creativity in learning English, and improves the quality of teaching. Corporate social media has become the most important way for corporate employees to record their lives, express opinions, share and communicate, and it is also one of the reliable and real-time sources of big data that reflects the true state of English learners. Real, accurate, and timely enterprise social media big data samples contain an enormous educational value, providing more possibilities for educational research. From the perspective of value, through sentiment analysis, topic mining, social network analysis, etc. on social media big data, learner portraits can be realized, thereby providing decision-making reference and support for stakeholders. This paper first builds a learning interest classification model based on TCNN-GRU deep learning, collects experimental data sets from an online English learner's social media platform and performs learning interest classification and labeling, and then uses the TCNN-GRU model to determine the user's learning interest tendency. On this basis, the concept of learning interest index is further proposed, and a neural network-based corporate social platform English learner portrait model is established. The experimental results show that, compared with the traditional machine learning model, convolutional neural network model, and recurrent neural network model, the TCNN-GRU model can obtain better results in learning interest classification. [ABSTRACT FROM AUTHOR]

It is crucial to ensure proper surveillance for the safety and security of people and their assets. The development of an aerial surveillance system might be very effective in catering to the challenges in surveillance systems. Current systems are expensive and complex. A cost-effective and efficient solution is required, which is easily accessible to anyone with a moderate budget. In aerial surveillance, quadcopters are equipped with state-of-the-art image processing technology that captures detailed photographs of every object underneath. A quadcopter-based solution is proposed to monitor desired premises for any unusual activities, like the movement of persons with weapons and face detection to achieve the desired surveillance. After detection of any unusual activity, the proposed system generates an alert for security personals. The proposed solution is based on quadcopter surveillance and video streaming for anomaly detection in the received video streams through deep learning models. A well-known FasterRCNN algorithm is modified for fast learning with feature reduction in the initial feature extraction step. Five different kinds of CNNs were evaluated for their ability to identify objects of interest in surveillance images. ResNet-50–based FasterRCNN with the highest average precision performed as an excellent solution for threat detection. The average precision of the system is 79% across all categories achieved. [ABSTRACT FROM AUTHOR]

The modelling and prediction of extreme temperature changes in enclosed compartments is a domain with applications ranging from residential fire alarms, industrial temperature sensors to search and rescue personnel safety systems. The spread of fire in enclosed compartments is a highly uncertain and nonlinear process. Hence, in safety-critical cases, any false negatives pose a serious threat to the safety of individuals such as firefighters that are engaged in rescue activities. This work aims to model the nonlinear fire spread behaviour as a temporal, deep learning-based regressive methodology. The objective is to efficiently identify abrupt and extreme temperature changes that often result in increases of 300+ °C. A major challenge in such time-series models is that of learning from historic time-series samples which are known to suffer from high noise levels, outliers and data imbalance. This work contributes on the development of a convolutional neural network (CNN)-long short-term memory (LSTM) methodology to handle temperature data originating from body-mounted and fixed sensors and develop a temperature increase warning mechanism. The main contribution exploits the contextualisation ability of CNN-LSTM to predict temperature changes in windows of 5–120 s. The model identifies the spatial temperature change patterns via a CNN encoder, which are then fed into an LSTM network. This regression mechanism is trained and validated against a set of unique fire spread conditions and involved live tests ranging from containers to residential and industrial units. The model's performance was evaluated with MAPE sensitivity analysis against data originating from body-mounted sensors and third-party NIST datasets. The outcome showed an error ranging from 0.89 to 2.05% to 5.46% and 6.23%, respectively. The model efficacy was also evaluated against a range of input–output temperature ranges from 5, 30 to 120-s windows and showed a FN rate of 2.15% for pre-alarm-to-normal and 3.11% for alarm-to-pre-alarm cases in body-mounted sensors and a higher FN rate of 5.14% reported for pre-alarm-to-normal case for the raised platfor sensor tests. [ABSTRACT FROM AUTHOR]