Background and Objectives: Coronary artery anomalies (CAAs) represent a group of rare cardiac abnormalities with an incidence of up to 1.2%. The aim of this retrospective study was to conduct a comprehensive epidemiological assessment of the prevalence of hypoplastic coronary arteries using coronary computed tomography angiography (CCTA) in patients with diagnosed CAAs and individuals presenting with cardiovascular manifestations in the north-eastern region of Romania. This study was motivated by the limited investigation of the CAAs conducted in this area. Methods: We analyzed data collected from 12,758 coronary computed tomography angiography (CCTA) records available at the “Prof. Dr. George I.M. Georgescu” Cardiovascular Diseases Institute, spanning the years 2012 to 2022. Results: Among 350 individuals with CAAs (2.7% of the total cohort), 71 patients (20.3% of the anomaly presenting group and 0.5% of the entire CCTA cohort) exhibited at least one hypoplastic coronary artery. The mean age of individuals diagnosed with hypoplastic coronary artery disease (HCAD) was 61 years, while the age distribution among them ranged from 22 to 84 years. Nearly equal cases of right and left dominance (33 and 31, respectively) were observed, with only 7 cases of co-dominance. Conclusions: HCAD may be considered underexplored in current published research, despite its potentially significant implications ranging to an increased risk of sudden cardiac arrest. The specific prevalence of HCAD among CAAs might be higher than previously reported, possibly reflecting better diagnostic accuracy of CCTA over classic coronary imaging. The absence of standard diagnostic and therapeutic protocols for HCAD underscores the necessity of a personalized approach for such cases. Full article

Evapotranspiration (ET) is a major component of the water budget in Bottomland Hardwood Forests (BHFs) and is driven by a complex intertwined suite of meteorological variables. The understanding of these interdependencies leading to seasonal variations in ET is crucial in better informing water resource management in the region. We used structural equation modeling and AIC modeling to analyze drivers of ET using Eddy covariance water flux data collected from a BHF located in the Russel Sage Wildlife Management Area (RSWMA). It consists of mature closed-canopy deciduous hardwood trees with an average canopy height of 27 m. A factor analysis was used to characterize the shared variance among drivers, and a path analysis was used to quantify the independent contributions of individual drivers. In our results, ET and net radiation (R_n) showed similar variability patterns with Vapor Pressure Deficit (VPD) and temperature in the spring, summer, and autumn seasons, while they differed in the winter season. The path analysis showed that R_n has the strongest influence on ET variations via direct and indirect pathways. In deciduous forests like BHFs, our results suggest that ET is more energy dependent during the growing season (spring and summer) and early non-growing season (autumn) and more temperature dependent during the winter season. Full article

Nuclear fusion is a potential source of energy that could supply the growing needs of the world population for millions of years. Several experimental thermonuclear fusion devices try to understand and control the nuclear fusion process. A very interesting diagnostic called Thomson scattering (TS) is performed in the Spanish fusion device TJ-II. This diagnostic takes images to measure the temperature and density profiles of the plasma, which is heated to very high temperatures to produce fusion plasma. Each image captures spectra of laser light scattered by the plasma under different conditions. Unfortunately, some images are corrupted by noise called stray light that affects the measurement of the profiles. In this work, we propose the use of deep learning models to reduce the stray light that appears in the diagnostic. The proposed approach utilizes a Pix2Pix neural network, which is an image-to-image translation based on a generative adversarial network (GAN). This network learns to translateimages affected by stray light to images without stray light. This allows for the effective removal of the noise that affects the measurements of the TS diagnostic, avoiding the need for manual image processing adjustments. The proposed method shows a better performance, reducing the noise up to $98\%$ inimages, which surpassesprevious works that obtained $85\%$ for the validation dataset. Full article

Point cloud registration is an important task in computer vision and robotics which is widely used in 3D reconstruction, target recognition, and other fields. At present, many registration methods based on deep learning have better registration accuracy in complete point cloud registration, but partial registration accuracy is poor. Therefore, a partial point cloud registration network, HALNet, is proposed. Firstly, a feature extraction network consisting mainly of adaptive graph convolution (AGConv), two-dimensional convolution, and convolution block attention (CBAM) is used to learn the features of the initial point cloud. Then the overlapping estimation is used to remove the non-overlapping points of the two point clouds, and the hybrid attention mechanism composed of self-attention and cross-attention is used to fuse the geometric information of the two point clouds. Finally, the rigid transformation is obtained by using the fully connected layer. Five methods with excellent registration performance were selected for comparison. Compared with SCANet, which has the best registration performance among the five methods, the RMSE(R) and MAE(R) of HALNet are reduced by 10.67% and 12.05%. In addition, the results of the ablation experiment verify that the hybrid attention mechanism and fully connected layer are conducive to improving registration performance. Full article

Globalization opens up economies and encourages the free movement of persons and factors of production. Diaspora investors and workers earn income in the process and make remittances to the migrating countries. We examine the impact of the remittance inflow on the macroeconomic performance of top emigrating countries, which comprise nine emerging and two advanced economies. We conduct group and individual country analyses with distinct econometric models (Feasible Quasi Generalized Least Squares and Dynamic Common Correlated Effects) using data between 1987 and 2021. The results reveal positive impact of remittance inflows on nominal GDP and nominal GDP per capita and on real GDP and real GDP per capita, although evidence on the latter is weaker. In all, the emigrating countries can benefit from diaspora remittance in terms of improved productivity and macroeconomic performance. We therefore recommend better systems to facilitate remittance receipt and policies to channel such flows more into investment activities. Full article

Abundant tuff mineral resources offer a promising solution to the shortage of fly ash (FA) and silica fume (SF) resources as emerging supplementary cementitious materials. However, a lack of clarity on its hydration mechanism has hindered its practical engineering application. In this study, high SiO₂-content tuff powder (TP) was examined to assess the mechanical and workability performance of mortar specimens with varying particle sizes of the TP as complete replacements for FA or SF. Microscopic analysis techniques, including X-ray diffraction (XRD), differential thermal analysis (DTG), and energy-dispersive X-ray spectroscopy (EDS), were employed to elucidate the hydration mechanism of the TP and its feasibility as a substitute for SF or FA. Results indicated that TP primarily functions as nuclei and filler, promoting cement hydration, with smaller particle sizes amplifying the hydration ability and increasing Ca(OH)₂ and C-S-H gel content. The specimens with TP (median particle size 7.58 μm) demonstrated 9.2% and 29.9% higher flexural and compressive strengths at 28 days, respectively, compared to the FA specimens of equal mass. However, fluidity decreased by 23.1% accordingly. Due to TP’s smaller specific surface area compared to SF, the TP specimens exhibited higher fluidity but with decreased strength relative to the SF specimens. Overall, TP shows potential as a replacement for FA with additional measures to ensure workability. Full article

Musculoskeletal injuries such as equine osteoarthritis, osteoarticular defects, tendonitis/desmitis, and muscular disorders are prevalent among sport horses, with a fair prognosis for returning to exercise or previous performance levels. The field of equine medicine has witnessed rapid and fruitful development, resulting in a diverse range of therapeutic options for musculoskeletal problems. Staying abreast of these advancements can be challenging, prompting the need for a comprehensive review of commonly used and recent treatments. The aim is to compile current therapeutic options for managing these injuries, spanning from simple to complex physiotherapy techniques, conservative treatments including steroidal and non-steroidal anti-inflammatory drugs, hyaluronic acid, polysulfated glycosaminoglycans, pentosan polysulfate, and polyacrylamides, to promising regenerative therapies such as hemoderivatives and stem cell-based therapies. Each therapeutic modality is scrutinized for its benefits, limitations, and potential synergistic actions to facilitate their most effective application for the intended healing/regeneration of the injured tissue/organ and subsequent patient recovery. While stem cell-based therapies have emerged as particularly promising for equine musculoskeletal injuries, a multidisciplinary approach is underscored throughout the discussion, emphasizing the importance of considering various therapeutic modalities in tandem. Full article

This study used DC magnetron sputtering technology to fabricate Sm-Fe films and systematically investigated the phase transition behavior of Sm-Fe films with different Fe ratios. It was found that at higher Fe content, the films consisted of Sm₂Fe₁₇ or SmFe₇ phases; as Fe content decreased, the films were mainly composed of SmFe₃ or SmFe₂ phases; at higher Sm content, the films primarily consisted of Sm phase. Sm is prone to volatilization at high temperatures, so Ta was used as a capping layer to effectively suppress Sm volatilization, successfully synthesizing pure SmFe₂ phase films at a nearly 1:2 ratio. The magnetic properties and magnetostrictive behavior of the SmFe₂ films were investigated, revealing that pure-phase SmFe₂ films exhibit good perpendicular magnetic anisotropy and magnetostriction properties. The larger stress along the perpendicular-to-film direction, resulting from the absence of substrate-induced constraints, contributes to the excellent perpendicular magnetic anisotropy of the films. This study successfully synthesized pure-phase SmFe₂ films and discovered a new method for fabricating perpendicularly anisotropic films. The research findings are of great significance for the efficient synthesis of desired films with high phase formation temperatures containing volatile elements. Full article

The partial discharge (PD) event in high-voltage insulation releases energy, exerts mechanical pressure, and generates elastic waves. Detecting and locating these PD events through short-duration acoustic pulses is well established, particularly in gas-insulated systems and oil-insulated transformers. However, its full potential remains untapped in solid insulation systems, where the propagation capability of the acoustic pulse and the acoustic reflections pose fundamental challenges to the acoustic emission (AE) detection technique. This study investigates the influence of reflections and multiple paths on the propagating acoustic pulse in polymeric insulating materials using a finite element method (FEM) in COMSOL. It was observed that the reflections from the boundary influence the propagating pulse’s shape, peak magnitude, and arrival time. An analytical MATLAB model further quantifies the impact of multiple propagation paths on the shape, magnitude, and arrival time of the pulse travelling in a cylinder. Additionally, a Perfect Matched Layer (PML) was implemented in the COMSOL model to eliminate the reflections from the boundary, and it revealed that the acoustic pulse magnitude decreases with distance following the inverse square law. In essence, the models aid in measuring how reflections contribute to the observed signals, facilitating the precise identification of the source of the PD event in the tested system. Full article

Bubble growth processes are highly complex processes, which are not only dependent on the foaming process parameters (temperature, pressure and blowing agent concentration) but also on the type and structure of the polymer used. Since the elongational viscosity at the bubble wall during bubble growth also depends on these influencing factors, the so-called transient elongational viscosity plays a key role in describing the gas bubble growth behavior in polymer melts. The model-based description of the transient elongational viscosity function is difficult due to its dependence on time, Hencky strain and strain rate. Therefore, representative viscosities or shear viscosity models are usually used in the literature to predict the bubble growth behavior. In this work, the transient equibiaxial elongational viscosity function at the bubble wall during bubble growth is described holistically for the first time. This is achieved by extending the so-called molecular stress function (MSF) model by superposition principles (temperature, pressure and blowing agent concentration) and by using the elongational deformation behavior (Hencky strain and strain rate) at the bubble wall during the initial, and thus viscosity-driven, bubble growth process. Therefore, transient uniaxial elongational viscosity measurements are performed and the non-linear MSF model parameters of the two investigated polymers PS (linear polymer chains) and PLA (long-chain branched polymer chains) are determined. By applying the superposition principles and by changing the strain mode parameter to the equibiaxial case in the MSF model, the transient equibiaxial viscosity master curve is obtained and used to describe the bubble growth process. The results show that the extended MSF model can fully predict the transient equibiaxial elongational viscosity function at the bubble wall during bubble growth processes. The bubble growth behavior over time can then be realistically described using the defined transient equibiaxial elongational viscosity function at the bubble wall. This is not possible, for example, with a representative viscosity and therefore clearly demonstrates the influence and importance of knowing the transient deformation behavior that prevails at the bubble wall during bubble growth processes. Full article

Based on the need for radiobiological databases, in this work, we mined experimental ionizing radiation data of human cells treated with X-rays, γ-rays, carbon ions, protons and α-particles, by manually searching the relevant literature in PubMed from 1980 until 2024. In order to calculate normal and tumor cell survival α and β coefficients of the linear quadratic (LQ) established model, as well as the initial values of the double-strand breaks (DSBs) in DNA, we used WebPlotDigitizer and Python programming language. We also produced complex DNA damage results through the fast Monte Carlo code MCDS in order to complete any missing data. The calculated α/β values are in good agreement with those valued reported in the literature, where α shows a relatively good association with linear energy transfer (LET), but not β. In general, a positive correlation between DSBs and LET was observed as far as the experimental values are concerned. Furthermore, we developed a biophysical prediction model by using machine learning, which showed a good performance for α, while it underscored LET as the most important feature for its prediction. In this study, we designed and developed the novel radiobiological ‘RadPhysBio’ database for the prediction of irradiated cell survival (α and β coefficients of the LQ model). The incorporation of machine learning and repair models increases the applicability of our results and the spectrum of potential users. Full article

Background and objectives: Hepatocellular carcinoma (HCC) is one of the most common malignancies worldwide and is caused by multiple factors. To explore novel targets for HCC treatment, we comprehensively analyzed the expression of HomeoboxB13 (HOXB13) and its role in HCC. Materials and Methods: The clinical significance of HCC was investigated using open gene expression databases, such as TIMER, UALCAN, KM, OSlihc, and LinkedOmics, and immunohistochemistry analysis. We also analyzed cell invasion and migration in HCC cell lines transfected with HOXB13-siRNA and their association with MMP9, E2F1, and MEIS1. Results: HOXB13 expression was higher in fibrolamellar carcinoma than in other histological subtypes. Its expression was associated with lymph node metastasis, histological stage, and tumor grade. It was positively correlated with immune cell infiltration of B cells (R = 0.246), macrophages (R = 0.182), myeloid dendritic cells (R = 0.247), neutrophils (R = 0.117), and CD4+ T cells (R = 0.258) and negatively correlated with immune cell infiltration of CD8+ T cells (R = −0.107). A positive correlation was observed between HOXB13, MMP9 (R = 0.176), E2F1 (R = 0.241), and MEIS1 (R = 0.189) expression (p < 0.001). The expression level of HOXB13 was significantly downregulated in both HepG2 and PLC/PFR/5 cell lines transfected with HOXB13-siRNA compared to that in cells transfected with NC siRNA (p < 0.05). Additionally, HOXB13 significantly affected cell viability and wound healing. Conclusions: HOXB13 overexpression may lead to poor prognosis in patients with HCC. Additional in vivo studies are required to improve our understanding of the biological role and the exact mechanism of action of HOXB13 in HCC. Full article

Background: Long-term rhythm monitoring (LTRM) can detect undiagnosed atrial fibrillation (AF) in patients at risk of AF and stroke. Circulating microRNAs (miRNAs), which have been shown to play a role in atrial electrical and structural remodelling, could help to select patients who would benefit most from LTRM. The aim of this study was to investigate whether patients with diabetes mellitus (DM) and hypertension and screen-detected subclinical AF (SCAF) using an insertable cardiac monitor (ICM) have significantly different plasma baseline levels of five selected miRNAs playing a role in the modulation of atrial electrical and structural remodelling (miR-21-5p, miR-29b-3p, miR-150-5p, miR-328-3p, and miR-432-5p) compared to those without SCAF. Methods: This study was performed at the outpatient clinic of a secondary academic teaching hospital between December 2013 and November 2015. Eligible patients were ≥65 years of age with DM and hypertension but without known heart diseases. All patients received an ICM. On the day of ICM implantation, blood samples for the measurement of plasma levels of the five miRNAs were drawn. In this post hoc analysis, we investigated their expression by reverse transcription-quantitative polymerase chain reaction. MiRNA plasma levels in patients with and without newly detected SCAF were compared. Results: We included 82 consecutive patients (median age of 71.3 years (IQR 67.4–75.1)), who were followed for a median of 588 days (IQR: 453–712 days). Seventeen patients (20.7%) had ICM-detected SCAF. Plasma levels of miR-328-3p, miR-29b-3p, miR-21-5p, miR-432-5p, and miR-150-5p were slightly but not significantly different in patients with incident SCAF compared with patients without. Conclusions: In patients with hypertension and DM, newly detected SCAF was not significantly associated with changes in expression levels of miR-21-5p, miR-29b-3p, miR-150-5p, miR-328-3p, and miR-432-5p. Full article

Brain tissue segmentation plays a critical role in the diagnosis, treatment, and study of brain diseases. Accurately identifying these boundaries is essential for improving segmentation accuracy. However, distinguishing boundaries between different brain tissues can be challenging, as they often overlap. Existing deep learning methods primarily calculate the overall segmentation results without adequately addressing local regions, leading to error propagation and mis-segmentation along boundaries. In this study, we propose a novel mis-segmentation-focused loss function based on a two-stage nnU-Net framework. Our approach aims to enhance the model’s ability to handle ambiguous boundaries and overlapping anatomical structures, thereby achieving more accurate brain tissue segmentation results. Specifically, the first stage targets the identification of mis-segmentation regions using a global loss function, while the second stage involves defining a mis-segmentation loss function to adaptively adjust the model, thus improving its capability to handle ambiguous boundaries and overlapping anatomical structures. Experimental evaluations on two datasets demonstrate that our proposed method outperforms existing approaches both quantitatively and qualitatively. Full article

Traditional methods for assessing the cleanliness of liquid metal are characterized by prolonged detection times, delays, and susceptibility to variations in sampling conditions. To address these limitations, an online cleanliness-analyzing system grounded in the method of the electrical sensing zone has been developed. This system facilitates real-time, in situ, and quantitative analysis of inclusion size and amount in liquid metal. Comprising pneumatic, embedded, and host computer modules, the system supports the continuous, online evaluation of metal cleanliness across various metallurgical processes in high-temperature environments. Tests conducted with gallium liquid at 90 °C and aluminum melt at 800 °C have validated the system’s ability to precisely and quantitatively detect inclusions in molten metal in real time. The detection procedure is stable and reliable, offering immediate data feedback that effectively captures fluctuations in inclusion amount, thereby meeting the metallurgical industry’s demand for real-time analyzing and control of inclusion cleanliness in liquid metal. Additionally, the system was used to analyze inclusion size distribution during the hot-dip galvanizing process. At a zinc melt temperature of 500 °C, it achieved a detection limit of 21 μm, simultaneously providing real-time data on the size and amount distribution of inclusions. This represents a novel strategy for the online monitoring and quality control of zinc slag throughout the hot-dip galvanizing process. Full article

Research on wildlife monitoring methods is a crucial tool for the conservation of rare wildlife in China. However, the fact that rare wildlife monitoring images in field scenes are easily affected by complex scene information, poorly illuminated, obscured, and blurred limits their use. This often results in unstable recognition and low accuracy levels. To address this issue, this paper proposes a novel wildlife identification model for rare animals in Giant Panda National Park (GPNP). We redesigned the C3 module of YOLOv5 using NAMAttention and the MemoryEfficientMish activation function to decrease the weight of field scene features. Additionally, we integrated the WIoU boundary loss function to mitigate the influence of low-quality images during training, resulting in the development of the NMW-YOLOv5 model. Our model achieved 97.3% for mAP50 and 83.3% for mAP50:95 in the LoTE-Animal dataset. When comparing the model with some classical YOLO models for the purpose of conducting comparison experiments, it surpasses the current best-performing model by 1.6% for mAP50:95, showcasing a high level of recognition accuracy. In the generalization ability test, the model has a low error rate for most rare wildlife species and is generally able to identify wildlife in the wild environment of the GPNP with greater accuracy. It has been demonstrated that NMW-YOLOv5 significantly enhances wildlife recognition accuracy in field environments by eliminating irrelevant features and extracting deep, effective features. Furthermore, it exhibits strong detection and recognition capabilities for rare wildlife in GPNP field environments. This could offer a new and effective tool for rare wildlife monitoring in GPNP. Full article

Detecting cracks in the pavement is a vital component of ensuring road safety. Since manual identification of these cracks can be time-consuming, an automated method is needed to speed up this process. However, creating such a system is challenging due to factors including crack variability, variations in pavement materials, and the occurrence of miscellaneous objects and anomalies on the pavement. Motivated by the latest progress in deep learning applied to computer vision, we propose an effective U-Net-shaped model named DepthCrackNet. Our model employs the Double Convolution Encoder (DCE), composed of a sequence of convolution layers, for robust feature extraction while keeping parameters optimally efficient. We have incorporated the TriInput Multi-Head Spatial Attention (TMSA) module into our model; in this module, each head operates independently, capturing various spatial relationships and boosting the extraction of rich contextual information. Furthermore, DepthCrackNet employs the Spatial Depth Enhancer (SDE) module, specifically designed to augment the feature extraction capabilities of our segmentation model. The performance of the DepthCrackNet was evaluated on two public crack datasets: Crack500 and DeepCrack. In our experimental studies, the network achieved mIoU scores of 77.0% and 83.9% with the Crack500 and DeepCrack datasets, respectively. Full article

We recently found that polyvinylpyrrolidone (PVP)-protected metal nanoparticles dispersed in water/butanol mixture spontaneously float to the air/water interface and form two-dimensional assemblies due to classical surface excess theory and Rayleigh–Bénard–Marangoni convection induced by butanol evaporation. In this study, we found that by leveraging this principle, a unique structure is formed where hetero gold nanospheres (AuNPs)/gold nanostars (AuNSs) complexes are dispersed within AuNP two-dimensional assemblies, obtained from a mixture of polyvinylpyrrolidone-protected AuNPs and AuNSs that interact electrostatically with the AuNPs. These structures were believed to form as a result of AuNPs/AuNSs complexes formed in the water/butanol mixture floating to the air/water interface and being incorporated into the growth of AuNP two-dimensional assemblies. These structures were obtained by optimizing the amount of mixed AuNSs, with excessive addition resulting in the formation of random three-dimensional network structures. The AuNP assemblies dispersed with AuNPs/AuNSs complexes exhibited significantly higher Raman (surface-enhanced resonance Raman scattering: SERRS) activity compared to simple AuNP assemblies, while the three-dimensional network structure did not show significant SERRS activity enhancement. These results demonstrate the excellent SERRS activity of AuNP two-dimensional assemblies dispersed with hetero AuNPs/AuNSs complexes. Full article

Asthma is a prevalent respiratory disease. The present study is designed to determine whether gut microbiota-derived tryptophan metabolites alleviate allergic asthma inflammation in ovalbumin (OVA)-induced mice and explore the effect and potential mechanism therein. Asthma model mice were constructed by OVA treatment, and kynurenine (KYN), indole-3-lactic acid (ILA), in-dole-3-carbaldehyde (I3C), and indole acetic acid (IAA) were administered by intraperitoneal injection. The percent survival, weight and asthma symptom score of mice were recorded. The total immunoglobulin E and OVA-specific (s)IgE in the serum and the inflammatory cytokines in the bronchoalveolar lavage fluid (BALF) were detected by the corresponding ELISA kits. The composition of the gut microbiota and tryptophan-targeted metabolism in mouse feces were analyzed using 16S rRNA gene sequencing and targeted metabolomics, respectively. The four tryptophan metabolites improved the percent survival, weight and asthma symptoms of mice, and reduced the inflammatory cells in lung tissues, especially I3C. I3C and IAA significantly (p < 0.05) downregulated the levels of OVA-IgE and inflammatory cytokines. KYN was observed to help restore gut microbiota diversity. Additionally, I3C, KYN, and ILA increased the relative abundance of Anaeroplasma, Akkermansia, and Ruminococcus_1, respectively, which were connected with tryptophan metabolic pathways. IAA also enhanced capability of tryptophan metabolism by the gut microbiota, restoring tryptophan metabolism and increasing production of other tryptophan metabolites. These findings suggest that tryptophan metabolites may modulate asthma through the gut microbiota, offering potential benefits for clinical asthma management. Full article

The wheat scab caused by Fusarium graminearum (F. graminearum) has seriously affected the yield and quality of wheat in China. In this study, gallic acid (GA), a natural polyphenol, was used to synthesize three azole-modified gallic acid derivatives (AGAs1–3). The antifungal activity of GA and its derivatives against F. graminearum was studied through mycelial growth rate experiments and field efficacy experiments. The results of the mycelial growth rate test showed that the EC₅₀ of AGAs–2 was 0.49 mg/mL, and that of AGAs–3 was 0.42 mg/mL. The biological activity of AGAs–3 on F. graminearum is significantly better than that of GA. The results of field efficacy tests showed that AGAs–2 and AGAs–3 significantly reduced the incidence rate and disease index of wheat scab, and the control effect reached 68.86% and 72.11%, respectively. In addition, preliminary investigation was performed on the possible interaction between AGAs–3 and F. graminearum using density functional theory (DFT). These results indicate that compound AGAs–3, because of its characteristic of imidazolium salts, has potential for use as a green and environmentally friendly plant-derived antifungal agent for plant pathogenic fungi. Full article

Cyberbullying is a serious problem in online communication. It is important to find effective ways to detect cyberbullying content to make online environments safer. In this paper, we investigated the identification of cyberbullying contents from the Bangla and Chittagonian languages, which are both low-resource languages, with the latter being an extremely low-resource language. In the study, we used both traditional baseline machine learning methods, as well as a wide suite of deep learning methods especially focusing on hybrid networks and transformer-based multilingual models. For the data, we collected over 5000 both Bangla and Chittagonian text samples from social media. Krippendorff’s alpha and Cohen’s kappa were used to measure the reliability of the dataset annotations. Traditional machine learning methods used in this research achieved accuracies ranging from 0.63 to 0.711, with SVM emerging as the top performer. Furthermore, employing ensemble models such as Bagging with 0.70 accuracy, Boosting with 0.69 accuracy, and Voting with 0.72 accuracy yielded promising results. In contrast, deep learning models, notably CNN, achieved accuracies ranging from 0.69 to 0.811, thus outperforming traditional ML approaches, with CNN exhibiting the highest accuracy. We also proposed a series of hybrid network-based models, including BiLSTM+GRU with an accuracy of 0.799, CNN+LSTM with 0.801 accuracy, CNN+BiLSTM with 0.78 accuracy, and CNN+GRU with 0.804 accuracy. Notably, the most complex model, (CNN+LSTM)+BiLSTM, attained an accuracy of 0.82, thus showcasing the efficacy of hybrid architectures. Furthermore, we explored transformer-based models, such as XLM-Roberta with 0.841 accuracy, Bangla BERT with 0.822 accuracy, Multilingual BERT with 0.821 accuracy, BERT with 0.82 accuracy, and Bangla ELECTRA with 0.785 accuracy, which showed significantly enhanced accuracy levels. Our analysis demonstrates that deep learning methods can be highly effective in addressing the pervasive issue of cyberbullying in several different linguistic contexts. We show that transformer models can efficiently circumvent the language dependence problem that plagues conventional transfer learning methods. Our findings suggest that hybrid approaches and transformer-based embeddings can effectively tackle the problem of cyberbullying across online platforms. Full article