Terrorist groups like ISIS have made considerable use of the dark web (DW) to carry out their malicious objectives such as spreading propaganda, recruiting and radicalizing new recruits, and secretly raising finances. Al-Hayat Media Center, an ISIS media agency, released a link on their forum describing how to access their DW website. It also sent out the identical message over Telegram, which included links to a Tor service with a “.onion” address. This study develops an analytical framework for scraping and analyzing the DW on the internet. The authors successfully tested a web crawler to collect account information for thousands of merchants and their related marketplace listings using a case study marketplace. The paper explains how to scrape DW marketplaces in the most viable and effective way possible. The findings of the case study support the validity of the proposed analytical framework, which is useful for academics researching this emerging phenomena as well as investigators looking into illegal behavior on the DW. Request access from your librarian to read this article's full text.

Multimodal emotion recognition is a developing field that analyzes emotions through various channels, mainly audio, video, and text. However, existing state-of-the-art systems focus on two to three modalities at the most, utilize traditional techniques, fail to consider emotional interplay, lack the scope to add more modalities, and aren’t efficient in predicting emotions accurately. This research proposes a novel approach using rule-based systems to convert non-verbal cues to text, inspired by a limited prior attempt, which lacked proper benchmarking. It achieves efficient multimodal emotion recognition by utilizing distilRoBERTa, a large language model, fine-tuned with a combined textual representation of audio (such as loudness, spectral flux, MFCCs, pitch stability, and emphasis) and visual features (action units) extracted from videos. This approach is evaluated using datasets RAVDESS and BAUM-1. It achieves high accuracy (93.18% in RAVDESS and 93.69% in BAUM-1) on both datasets, performing on par with the SOTA (State-of-the-Art) systems, if not slightly better. Furthermore, the research highlights the potential for incorporating additional modalities by transforming them into text using rule-based systems and utilizing them to refine further pre-trained large language models giving rise to a more comprehensive approach to emotion recognition.

The tremendous growth in online activity and the Internet of Things (IoT) led to an increase in cyberattacks. Malware infiltrated at least one device in almost every household. Various malware detection methods that use shallow or deep IoT techniques were discovered in recent years. Deep learning models with a visualization method are the most commonly and popularly used strategy in most works. This method has the benefit of automatically extracting features, requiring less technical expertise, and using fewer resources during data processing. Training deep learning models that generalize effectively without overfitting is not feasible or appropriate with large datasets and complex architectures. In this paper, a novel ensemble model, Stacked Ensemble—autoencoder, GRU, and MLP or SE-AGM, composed of three light-weight neural network models—autoencoder, GRU, and MLP—that is trained on the 25 essential and encoded extracted features of the benchmark MalImg dataset for classification was proposed. The GRU model was tested for its suitability in malware detection due to its lesser usage in this domain. The proposed model used a concise set of malware features for training and classifying the malware classes, which reduced the time and resource consumption in comparison to other existing models. The novelty lies in the stacked ensemble method where the output of one intermediate model works as input for the next model, thereby refining the features as compared to the general notion of an ensemble approach. Inspiration was drawn from earlier image-based malware detection works and transfer learning ideas. To extract features from the MalImg dataset, a CNN-based transfer learning model that was trained from scratch on domain data was used. Data augmentation was an important step in the image processing stage to investigate its effect on classifying grayscale malware images in the MalImg dataset. SE-AGM outperformed existing approaches on the benchmark MalImg dataset with an average accuracy of 99.43%, demonstrating that our method was on par with or even surpassed them.

This research article proposes a federated learning-based intrusion detection methodology that leverages Enhanced Ghost_BiNet, a novel deep learning model, to enhance the security of information sharing and detection accuracy. Federated learning, a privacy-preserving machine learning technique, is utilized to enable multiple entities to collaboratively train a global intrusion detection model without sharing sensitive data. The proposed system first trains local models using Enhanced Ghost_BiNet, which integrates GhostNet and Bidirectional Gated Recurrent Unit (BiGRU). To optimize the model’s performance, the Chaotic Chebyshev Artificial Humming Bird (CAh) algorithm is employed. Homomorphic encryption is applied to encrypt the local model updates, enhancing data privacy and security. Server-side aggregation of updates and collaborative optimization are introduced to minimize communication rounds during data aggregation. The results demonstrate that the Enhanced Ghost_BiNet outperforms traditional models like GhostNet, BiGRU, RNN, Auto Encoder, and CNN in terms of accuracy, precision, recall, F-Score, and mean square error (MSE). For instance, the Enhanced Ghost_BiNet achieves an accuracy of 99.24% on the KDD CUP 99 dataset, surpassing the other models by a significant margin. The proposed methodology provides a robust and secure approach to intrusion detection, ensuring the confidentiality of sensitive data while improving detection accuracy.

In this paper, a novel Elliptic Crypt with Secured Blockchain-backed Federated Q-Learning Framework is proposed to offer an intelligent healthcare system that mitigates the attacks and data misused by malicious intruders. Initially, the entered IoMT data is collected from publicly available datasets and encrypted using the Extended Elliptic Curve Cryptography (E_ECurCrypt) technique for ensuring the security. This encrypted data is fed as an input to the blockchain-powered collaborative learning model. Here, the federated Q-learning model trains the inputs and analyzes the presented attacks to ensure better privacy protection. Afterwards, the data is securely stored in decentralized blockchain technology. Subsequently, an effective Delegated Proof of Stake (Del_PoS) consensus algorithm is used to validate the proposed framework. The experiment is conducted using the WUSTL-EHMS-2020 dataset and the performances are analyzed by evaluating multiple matrices and compared to other existing methods. The performance of the proposed framework can be assessed using multiple matrices and the results will be compared to other existing methods. As a result, the proposed method has achieved 99.23% accuracy, 98.42% precision, 98.12% recall, 98.27% F1 score, 59080.506 average throughput, 59080.506 average decryption time 1.94 seconds and an average encryption time of 1.84 seconds and are superior to conventional methods.

Adverse Drug Reactions(ADRs) due to drug-drug interactions present a public health problem worldwide that deserves attention due to its impact on mortality, morbidity and healthcare costs. They have been a major challenge in healthcare with the ever increasing complexity of therapeutics and an aging population in many regions. At present, no standard method to detect such adverse drug reactions exist until otherwise reported by patients after the drug is released to the market. Further, several studies show that it is extremely challenging to detect these rare cases during clinical trials held before the drug is released. Therefore, a reliable and efficient technique to predict such side effects before the release of the drug to the market is the need of the hour. Through the power of Graph Neural Networks and the knowledge representation abilities of self supervised learning, we designed an effective framework to model drug-drug interactions by leveraging the spatial and physical properties of drugs by representing them as molecular graphs. Through this approach, we developed a technique that resembles the dynamics of a chemical interaction. On training and testing this approach on the TwoSIDES Polypharmacy Dataset by Theraputic Data Commons(TDC), we achieve state of the art results by obtaining a precision of 75% and accuracy of 90% on the test dataset. Further, we also perform a case study on the DrugBank dataset and compare our results on the interaction type prediction task in-order to validate our approach on the drug-drug interaction domain and achieve excellent results with a precision, F1 and accuracy of 99%. Our study and experimental approaches lays the groundwork for further research on side-effect prediction through drug-drug interaction and the use of Graph Neural Networks in the field of Molecular Biology.