Summary In the rapidly evolving landscape of fifth generation Internet of Things (5G‐IoT) networks, Device‐to‐Device (D2D) communication has emerged as a promising paradigm to enhance secrecy transmission rate (STR) and connectivity. However, the security of D2D communications in the presence of eavesdroppers remains a critical challenge. This article investigates the problem of optimizing D2D transmit power to achieve secure D2D communication while considering the presence of random eavesdroppers in 5G‐IoT networks. We propose a novel secrecy‐based power control approach (SRMWPCA) approach to model the random distribution of eavesdroppers in the network, taking into account their varying distances from D2D pairs and deliberately increasing interference at the eavesdropper's link. By leveraging tools from stochastic geometry, we derive an analytical expression for the secrecy transmission probability (STP), which quantifies the probability of eavesdroppers successfully decoding the D2D transmission. In this analysis, we have incorporated practical considerations such as channel fading, path loss, and interference from other devices. To enhance the security of D2D communication, we formulate an optimization problem to determine the optimal transmit power levels for D2D pairs, subject to constraints on the secrecy transmission probability and interference to the cellular network. We propose an efficient algorithm to find the power allocation that maximizes the secrecy outage performance while meeting these constraints. Simulation results demonstrate the effectiveness of the proposed approach in achieving secure D2D communication in 5G‐IoT networks with random eavesdroppers. The performance of the proposed SRMWPCA approach improved by 23.25% and 20.9% compared with standard approaches in terms of the secrecy rate and throughput of the users from malicious attacks.

Summary The wireless sensor network (WSN) has been assimilated into modern society and is utilized in many crucial application domains, including animal monitoring, border surveillance, asset monitoring, and so forth. These technologies aid in protecting the place of the event's occurrence from the adversary. Maintaining privacy concerning the source location is challenging due to the sensor nodes' limitations and efficient routing strategies. Hence, this research introduces a novel source location privacy preservation using the awareness‐based Gannet with random‐Dijkstra's algorithm (AGO‐RD). The network is initialized by splitting the hotspot and non‐hotspot region optimally using the proposed awareness‐based Gannet (AGO) algorithm. Here, the multi‐objective fitness function is utilized to initialize the network based on factors like throughput, energy consumption, latency, and entropy. Then, the information is forwarded to the phantom node in the non‐hotspot region to preserve the source location's privacy, which is far from the sink node. The proposed random‐Dijkstra algorithm is utilized to route the information from the phantom node to the sink with more security. Analysis of the proposed AGO‐RD‐based source location privacy preservation technique in terms of delay, throughput, network lifetime, and energy consumption accomplished the values of 6.52 ms, 95.68%, 7109.9 rounds, and 0.000125 μJ.

Detecting signs of suicidal thoughts on social media is paramount for preventing suicides, given the platforms' role as primary outlets for emotional expression. Traditional embedding techniques focus solely on semantic analysis and lack the sentiment analysis essential for capturing emotions. This limitation poses challenges in developing high-accuracy models. Additionally, previous studies often rely on a single dataset, further constraining their effectiveness. To overcome these challenges, this study proposes an innovative approach that integrates embedding techniques such as BERT, which offers semantic and syntactic analysis of the posts, with sentiment analysis provided by VADER scores extracted from the VADER sentiment analysis tool. The identified features are then input into the proposed optimised hybrid deep learning model, specifically the Bi-GRU and Attention incorporated with Stacked or Stacking Classifier (Decision Tree, Random Forest, Gradient Boost, as the base classifier and XGBoost as meta classifier), which undergoes optimisation using the grid search technique to enhance detection capabilities. In evaluations, the model achieved an impressive accuracy and F1-score of 98% on the Reddit dataset and 97% on the twitter (formally known as X) dataset. The research evaluates the efficacy of several machine learning models, encompassing Decision Trees, Random Forests, Gradient Boosting, and XGBoost. Moreover, it examines sophisticated models like LSTM with Attention, Bi-LSTM with Attention, and Bi-GRU with Attention, augmented with word embeddings such as BERT, MUSE, and fastText, alongside the fusion of sentiment VADER score. These results emphasise the promise of a holistic strategy that combines advanced feature embedding techniques with semantic features, showcasing a notably efficient detection of suicidal ideation on social media.

Summary The evolution of IoT, 5G and 6G aims to provide almost zero latency. Computation tasks size is different for different users. A framework for task computation achieving almost zero latency for stochastic demand is a challenge. A relay‐based D2D mobile edge computing (MEC) system is proposed. The idle device present in the networks is used as relay resources (RS). Mobile devices (MDs) communicate task to relay resources (RS) using D2D communication link. The RS perform computation and offloaded to edge server (ES). It aims to minimize total cost, energy expenditure and overall latency. Problem is formulated as mixed‐integer nonlinear‐constrained problem (MINCP). A three‐step algorithm to optimize relay selection, power allocation and computation resource allocation is proposed. In the initial step optimal relay selection is obtained by the Kuhn‐Munkres (KM) algorithm. In the next step, power allocation is obtained using Q‐learning. In the last step, the main problem is converted into a cost optimization problem deciphered by the proposed algorithm. The substantial simulation results indicate the relay‐based MEC system to achieve an astounding outcome in terms of latency, energy consumption and cost. Compared with other baseline methods, the proposed algorithm can achieve reduced energy consumption and cost for almost zero latency.

Summary Due to the remarkable development of Internet technologies, a great deal of valuable digital data is now transmitted over public networks. To guarantee the security of this data during the transfer process, the authentication of its integrity is extremely important. This paper introduces a robust and secure dual‐watermarking‐based fusion of watermarking, optimization, and a compression method utilizing non‐sub‐sampled contourlet transform (NSCT), redundant discrete wavelet transform (RDWT), and singular value decomposition (SVD). In our method, we first apply the NSCT to a higher entropy sub‐band of the host image. Then, our method uses RDWT‐SVD on higher frequency coefficients of the NSCT image. A similar procedure is followed for both mark images. Finally, an appropriate scaling factor, as obtained by fuzzy inference system, is used to invisibly embed the singular values of both mark data into the host image. Here, any more important mark data are scrambled before the embedding process. The simulation tests reveal that the proposed technique is not only imperceptible and secure but also robust against common attacks. The suggested method has a superior ability to extract hidden information than previous conventional techniques.