Browse Articles

The authors develop a new method for training machine learning algorithms to differentiate between hate speech and cultural speech in online platforms.

In our modern age, the burgeoning use of radios and radars has resulted in competition for electromagnetic spectrum resources. With recent research highlighting solutions to radio and radar mutual interference, there is a desperate need for a cost-effective configuration that permits a radar-radio joint system. In this study, the authors have set out to determine the feasibility of using single-tone continuous-wave radars in a radar-joint system. With this system, they aim to facilitate cost-effective near-field target detection by way of the popularized 2.4-GHz industrial, scientific, and medical (ISM) band.

Each year, over 100,000 patients die from Sudden Unexpected Death in Epilepsy (SUDEP). A reliable seizure warning system can help patients stay safe. This work presents a comprehensive, comparative analysis of three different signal processing algorithms for automated seizure/ictal detection. The experimental results show that the proposed methods can be effective for accurate automated seizure detection and monitoring in clinical care.

In the battle against Alzheimer's disease, early detection is critical to mitigating symptoms in patients. Here, the authors use a collection of MRI scans, layering with deep learning computer modeling, to investigate early stages of AD which can be hard to catch by human eye. Their model is successful, able to outperform previous models, and detected regions of interest in the brain for further consideration.

Staphylococcus aureus (S. aureus) has a mortality rate of up to 30% in developing countries. The purpose of this experiment was to determine if enzymatic and volatile compound-based approaches would perform more quickly in comparison to existing S. aureus diagnostic methods and to evaluate these novel methods on accuracy. Ultimately, this device provided results in less than 30 seconds, which is much quicker than existing methods that take anywhere from 10 minutes to 48 hours based on approach. Statistical analysis of accuracy provides preliminary confirmation that the device based on enzymatic and volatile compound-based approaches can be an accurate and time-efficient tool to detect pathogenic S. aureus.

Effective treatment of depression requires early detection. Depressive symptoms overlap with olfactory regions, which led to several studies of the correlation between sense of smell and depression. The alarming rise of depression, its related crimes, suicides, and lack of inexpensive, quick tools in detecting early depression — this study aims in demonstrating decreased olfaction and depression correlation. Forty-two subjects (ages 13-83) underwent POC-MON (Pocket Lemon) assessment — an oven-dried lemon peel sniff test, subjected to distance measurement when odor first detected (threshold) and completed Patient Health Questionnaires (PHQ-9). POC-MON and PHQ-9 scores yielded a correlation of 20% and 18% for the right and left nostrils, respectively. Among male (n=17) subjects, the average distance of POC-MON and PHQ-9 scores produced a correlation of 14% and 16% for the right and left nostrils, respectively. Females (n=25) demonstrated a correlation of 28% and 21% for the right and left nostrils, respectively. These results suggest the correlation between olfaction and depression in diagnosing its early-stage, using a quick, inexpensive, and patient-friendly tool — POC-MON.

Globally, the cultivation of 77.8 million tons of grapes each year underscores their significance in both diets and agriculture. However, grapevines face mounting threats from diseases such as black rot, Esca, and leaf blight. Traditional detection methods often lag, leading to reduced yields and poor fruit quality. To address this, authors used machine learning, specifically deep learning with Convolutional Neural Networks (CNNs), to enhance disease detection.

Although the 5-year survival rate for colorectal cancer is below 10%, it increases to greater than 90% if it is diagnosed early. We hypothesized from our research that analyzing non-synonymous single nucleotide variants (SNVs) in a patient's exome sequence would be an indicator for high genetic risk of developing colorectal cancer.

The purpose of this investigation is to develop a hydrogel to aid skin regeneration by creating an extracellular matrix for fibroblast growth with antibacterial and infection-detection properties. Authors developed two natural hydrogels based on pectin and potato peels and characterized the gels for fibroblast compatibility through rheology, scanning electron microscopy, swelling, degradation, and cell cytotoxicity assays. Overall, this experiment fabricated various hydrogels capable of acting as skin substitutes and counteracting infections to facilitate wound healing. Following further testing and validation, these hydrogels could help alleviate the 13-billion-dollar financial burden of foot ulcer treatment.

Osteosarcoma is a type of bone cancer that affects young adults and children. Early diagnosis of osteosarcoma is crucial to successful treatment. The current methods of diagnosis, which include imaging tests and biopsy, are time consuming and prone to human error. Hence, we used deep learning to extract patterns and detect osteosarcoma from histological images. We hypothesized that the combination of two different technologies (transfer learning and data augmentation) would improve the efficacy of osteosarcoma detection in histological images. The dataset used for the study consisted of histological images for osteosarcoma and was quite imbalanced as it contained very few images with tumors. Since transfer learning uses existing knowledge for the purpose of classification and detection, we hypothesized it would be proficient on such an imbalanced dataset. To further improve our learning, we used data augmentation to include variations in the dataset. We further evaluated the efficacy of different convolutional neural network models on this task. We obtained an accuracy of 91.18% using the transfer learning model MobileNetV2 as the base model with various geometric transformations, outperforming the state-of-the-art convolutional neural network based approach.