ABSTRACT
Cybercriminals often register many pornographic or gambling domains (known as abusive domains) with similar character compositions in bulk to reduce their investment in buying domains and make it easier for clients to remember and spread them. Therefore, this study combines the ideas of text similarity and text generation and proposes an abusive domain generation model based on GRU for rapidly generating new abusive domain names from known ones. Additionally, we develop a two-layer detection system for pornography and gambling domains using fastText and CNN models to obtain an abusive domain dataset for model training and validation. In the end, our detection system identifies pornographic and gambling domains with 99% precision while balancing correctness and speed. By inputting 40,000 random keywords into the abusive domain generation model, we obtained 130,220 online domains that served web pages, of which about 66% were pornographic or gambling domains. The results show that by exploiting cybercriminals' behaviors in registering abusive domain names, such as bulk registration of similar domain names, we can prospectively acquire a large number of new abusive domains based on known ones. This study demonstrates that predicting new abusive domains not only expands the domain blacklist but also allows researchers to target the generated suspicious domains and dispose of them in time before they show abusive behavior.
ABSTRACT
As Corona Virus Disease (COVID-19) pandemic strikes the world, retail industry has been severely impacted by staff shortage and high risk of virus outbreak. However, most of existing smart retail solutions is associated with high deployment and maintenance cost that are infeasible for small retail stores. As an effort to mitigate the issue, a computer vision-powered smart cashierless checkout system is proposed based on You Only Look Once (YOLO) v5 and MobileNet V3 for product recognition along with 3-stage image synthesis framework that includes crop and paste algorithm, GAN-based shadow synthesis and light variation algorithm. By using 3000 images generated from the framework, proposed model was trained and optimized with TensorRT. Experimental result shows that the lightweight model can be deployed on affordable edge devices like Jetson Nano while achieving high Mean Average Precision (mAP) of 98.2%, Checkout Accuracy (cAcc) of 89.17% with only 0.142s of inference time. © 2022, Success Culture Press. All rights reserved.
ABSTRACT
Peptides are increasingly regarded as promising antagonists to combat Severe Acute Respiratory Syndrome Coronavirus type 2 (SARS-CoV-2). In a recent computational study, we uncovered that mealworm proteins, following in silico gastrointestinal digestion, could be a promising source of peptides that potentially block the entry of SARS-CoV-2 into the host cells. In this study, we furthered our investigation to search for mealworm peptides that potentially target SARS-CoV-2 spike glycoprotein, main protease and papain-like protease. Among the 1588 peptide fragments screened, two peptides PKWF and VHRKCF stood out as putative multi-target peptides based on molecular docking analysis. Using in silico tools, we also predicted intermolecular interactions that allow binding of the peptides to the target proteins. Relative importance of the individual residues in the two sequences concerning binding stability to the target proteins was investigated. Physicochemical properties of the peptides were also predicted and discussed in relation to their binding to the targets. Overall, our findings suggest that PKWF and VHRKCF could be potential prophylactic or therapeutic agents against SARS-CoV-2. We hope that our findings could pave the way for and benefit future discovery of multi-target anti-SARS-CoV-2 agents from insect proteins, particularly from mealworms. © 2021 Malaysian Society for Biochemistry and Molecular Biology. All rights reserved.