Robust Epileptic Seizure Detection Using Long Short-Term Memory and Feature Fusion of Compressed Time-Frequency EEG Images.

Epilepsy is a prevalent neurological disorder with considerable risks, including physical impairment and irreversible brain damage from seizures. Given these challenges, the urgency for prompt and accurate seizure detection cannot be overstated. Traditionally, experts have relied on manual EEG signal analyses for seizure detection, which is labor-intensive and prone to human error. Recognizing this limitation, the rise in deep learning methods has been heralded as a promising avenue, offering more refined diagnostic precision. On the other hand, the prevailing challenge in many models is their constrained emphasis on specific domains, potentially diminishing their robustness and precision in complex real-world environments. This paper presents a novel model that seamlessly integrates the salient features from the time-frequency domain along with pivotal statistical attributes derived from EEG signals. This fusion process involves the integration of essential statistics, including the mean, median, and variance, combined with the rich data from compressed time-frequency (CWT) images processed using autoencoders. This multidimensional feature set provides a robust foundation for subsequent analytic steps. A long short-term memory (LSTM) network, meticulously optimized for the renowned Bonn Epilepsy dataset, was used to enhance the capability of the proposed model. Preliminary evaluations underscore the prowess of the proposed model: a remarkable 100% accuracy in most of the binary classifications, exceeding 95% accuracy in three-class and four-class challenges, and a commendable rate, exceeding 93.5% for the five-class classification.

AI-Enabled Traffic Control Prioritization in Software-Defined IoT Networks for Smart Agriculture.

Smart agricultural systems have received a great deal of interest in recent years because of their potential for improving the efficiency and productivity of farming practices. These systems gather and analyze environmental data such as temperature, soil moisture, humidity, etc., using sensor networks and Internet of Things (IoT) devices. This information can then be utilized to improve crop growth, identify plant illnesses, and minimize water usage. However, dealing with data complexity and dynamism can be difficult when using traditional processing methods. As a solution to this, we offer a novel framework that combines Machine Learning (ML) with a Reinforcement Learning (RL) algorithm to optimize traffic routing inside Software-Defined Networks (SDN) through traffic classifications. ML models such as Logistic Regression (LR), Random Forest (RF), k-nearest Neighbours (KNN), Support Vector Machines (SVM), Naive Bayes (NB), and Decision Trees (DT) are used to categorize data traffic into emergency, normal, and on-demand. The basic version of RL, i.e., the Q-learning (QL) algorithm, is utilized alongside the SDN paradigm to optimize routing based on traffic classes. It is worth mentioning that RF and DT outperform the other ML models in terms of accuracy. Our results illustrate the importance of the suggested technique in optimizing traffic routing in SDN environments. Integrating ML-based data classification with the QL method improves resource allocation, reduces latency, and improves the delivery of emergency traffic. The versatility of SDN facilitates the adaption of routing algorithms depending on real-time changes in network circumstances and traffic characteristics.

A Deep Learning-Based Intrusion Detection System for MQTT Enabled IoT.

A large number of smart devices in Internet of Things (IoT) environments communicate via different messaging protocols. Message Queuing Telemetry Transport (MQTT) is a widely used publish-subscribe-based protocol for the communication of sensor or event data. The publish-subscribe strategy makes it more attractive for intruders and thus increases the number of possible attacks over MQTT. In this paper, we proposed a Deep Neural Network (DNN) for intrusion detection in the MQTT-based protocol and also compared its performance with other traditional machine learning (ML) algorithms, such as a Naive Bayes (NB), Random Forest (RF), k-Nearest Neighbour (kNN), Decision Tree (DT), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRUs). The performance is proved using two different publicly available datasets, including (1) MQTT-IoT-IDS2020 and (2) a dataset with three different types of attacks, such as Man in the Middle (MitM), Intrusion in the network, and Denial of Services (DoS). The MQTT-IoT-IDS2020 contains three abstract-level features, including Uni-Flow, Bi-Flow, and Packet-Flow. The results for the first dataset and binary classification show that the DNN-based model achieved 99.92%, 99.75%, and 94.94% accuracies for Uni-flow, Bi-flow, and Packet-flow, respectively. However, in the case of multi-label classification, these accuracies reduced to 97.08%, 98.12%, and 90.79%, respectively. On the other hand, the proposed DNN model attains the highest accuracy of 97.13% against LSTM and GRUs for the second dataset.

CAFD: Context-Aware Fault Diagnostic Scheme towards Sensor Faults Utilizing Machine Learning.

Sensors' existence as a key component of Cyber-Physical Systems makes it susceptible to failures due to complex environments, low-quality production, and aging. When defective, sensors either stop communicating or convey incorrect information. These unsteady situations threaten the safety, economy, and reliability of a system. The objective of this study is to construct a lightweight machine learning-based fault detection and diagnostic system within the limited energy resources, memory, and computation of a Wireless Sensor Network (WSN). In this paper, a Context-Aware Fault Diagnostic (CAFD) scheme is proposed based on an ensemble learning algorithm called Extra-Trees. To evaluate the performance of the proposed scheme, a realistic WSN scenario composed of humidity and temperature sensor observations is replicated with extreme low-intensity faults. Six commonly occurring types of sensor fault are considered: drift, hard-over/bias, spike, erratic/precision degradation, stuck, and data-loss. The proposed CAFD scheme reveals the ability to accurately detect and diagnose low-intensity sensor faults in a timely manner. Moreover, the efficiency of the Extra-Trees algorithm in terms of diagnostic accuracy, F1-score, ROC-AUC, and training time is demonstrated by comparison with cutting-edge machine learning algorithms: a Support Vector Machine and a Neural Network.

Frequency and causes of bilateral occular trauma.

OBJECTIVE: To determine the frequency and causes of bilateral ocular trauma. DESIGN: A descriptive case series. PLACE AND DURATION OF STUDY: Khyber Institute of Ophthalmic Medical Sciences, Hayatabad Medical Complex, Peshawar from October 1999 to September 2006. PATIENTS AND METHODS: All patients coming to the hospital with bilateral eye trauma and requiring admission were recruited into the study. The details of patients' demographics, risk factors, ocular examination, treatment offered and final visual acuity were noted and described as frequency and percentages. RESULTS: Out of a total of 1551 patients of hospitalized ocular trauma, 46 (2.9%, 92 eyes) had bilateral ocular trauma. The majority (54.3%) were due to landmine blast injuries followed by dynamite blast in 10.8%, coalmine blast and firearm injury in 6.5% each. Pressure cooker explosion and road traffic accident was the cause in 4.3% each. Gas cylinder and automobile battery explosion, alkali and acid burn, assault and incidental trauma occurred in 2.1%. Sixty three percent were between 16 and 40 years of age. Males were affected in 93.4%. Corneal and / or scleral repair was done in 58.6%, conjunctival and or corneal foreign body removal in 26% and extracapular cataract extraction with intraocular lens implantation in 16.3%. The visual acuity was in the range of 6/60 and perception of light in 54.3%, while in 21.7%, there was no perception of light at the time of admission. Due to severity of injury, the final visual acuity was poor and only 28.2% regained vision between 6/18 and 6/60. CONCLUSION: In this series, landmine, dynamite and coalmine blasts were the major causes of bilateral ocular trauma. Victims were usually young males. Due to severity of ocular trauma, majority had poor visual outcome.