Processing Techniques Research Articles

Spatial and temporal deep learning algorithms for defect segmentation in infrared thermographic imaging of carbon fibre-reinforced polymers

ABSTRACT For non-destructive evaluation, the segmentation of infrared thermographic images of carbon fibre composites is a critical task in material characterisation and quality assessment. This paper presents a study on the application of image processing techniques, particularly adaptive thresholding, and advanced neural network models, including U-Net, DeepLabv3, and BiLSTM, for the segmentation of infrared images. This work introduces the innovative combination of DeepLabv3 and BiLSTM applied in infrared images of carbon fibre-reinforced polymer samples for the first time, proposing it as a novel approach for enhancing the accuracy of segmentation tasks. An experimental comparison of these models was conducted to assess their effectiveness in identifying artificial defects in these images. The performance of each model was evaluated using the F1-Score and Intersection over Union (IoU) metrics. The results demonstrate that the proposed combination of DeepLabv3 and BiLSTM outperforms other methods, achieving an F1-Score of 0.96 and an IoU of 0.83, showcasing its potential for advanced material analysis and quality control.

A novel method for multi-matrix arsenic speciation analysis by anion-exchange HPLC-ICP-MS in the framework of the third (French) total diet study.

This study presents the development and validation of a precise analytical method for the speciation analysis of arsenic (As) compounds, including inorganic species [As(III) and As(V)] and organic species such as monomethylarsonic acid (MMA) and dimethylarsinic acid (DMA). The method employs anion-exchange high-performance liquid chromatography (AE HPLC) coupled with inductively coupled plasma-mass spectrometry (ICP-MS). To optimize the sample preparation process, microwave-assisted extraction (MAE) and heat-assisted extraction (HAE) techniques were evaluated and compared. Separation of all four arsenic species was achieved with baseline resolution within 10 min, utilizing an anion-exchange column and a mobile phase gradient of 0.5 and 5 mM ammonium carbonate with 3% (v/v) methanol at a pH of 9.3. The method validation followed the accuracy profile approach, employing six different food matrices analyzed in duplicate over six separate days within a 6-week period. The method demonstrated good intermediate reproducibility, with coefficients of variation (CVR) ranging from 4.7% to 5.5%, while the bias was < 3%. The limit of quantification (LOQ) for the four species was 6.25 μg/kg (dry weight, dw), and the limit of detection (LOD) was 1.88 µg/kg (dw). These results confirm the robustness, accuracy, and suitability of the method for routine As speciation analysis across a variety of food matrices. The method is specifically intended to be applied to the analysis of a panel of food samples as part of the ongoing (French) third total diet study.

Operando Synchrotron X-Ray Absorption Spectroscopy: A Key Tool for Cathode Material Studies in Next-Generation Batteries.

Rechargeable batteries are central to modern energy storage systems, from portable electronics to electric vehicles. The cathode material, a critical component, largely dictates a battery's energy density, capacity, and overall performance. This review focuses on the application of operando X-ray absorption spectroscopy (XAS) to study cathode materials in Li-ion, Na-ion, Li-S, and Na-S batteries. Operando XAS provides real-time insights into the local electronic structure, oxidation states, and coordination environments, which are crucial for understanding complex electrochemical processes, such as redox reactions, phase transitions, and structural degradation. The review highlights the strengths of hard and soft XAS techniques in probing transition metal (TM) and anionic redox processes, particularly in layered oxide cathodes, where reversible oxygen redox and TM behavior are pivotal. The role of operando XAS is also explored in analyzing conversion-type S-based cathodes, where it helps unravel the intricate reaction mechanisms. Furthermore, the review addresses the challenges of in situ cell design for operando XAS, especially under ultrahigh vacuum conditions for soft XAS. By discussing recent advancements and future directions, this review underscores the critical role of operando XAS in driving innovation and improving the design and performance of next-generation battery technologies.

Corporate sponsorship as cultural policy: tax incentives in Brazilian contemporary art

ABSTRACT Unprecedented corporate intervention and the withdrawal of government support from the arts and culture characterized the 1980s. Driven by the Thatcher and Reagan administrations, which promoted corporate patronage and art sponsorship, this trend shaped cultural policies worldwide and was adopted by numerous countries. Following this global movement, Brazil introduced tax incentives in the early 1990s to encourage private companies to fund cultural events and institutions. While the literature on Brazilian cultural tax incentives is extensive, researchers have overlooked the connection between these incentives and the country’s artistic production fields. Drawing on data from the past decade of cultural funding and applying Natural Language Processing and machine learning techniques to isolate sponsorship patterns related to contemporary art, this paper examines the financial landscape of Brazil’s contemporary art circuit. The findings reveal that inequalities in resource distribution for contemporary art sponsorship exceed those in other fields, such as music and theater, highlighting the highly institutionalized nature of contemporary art, its reliance on corporate sponsorship, and the substantial costs associated with its production.

Design, development and performance of a Fe-Mn-Si-Cu alloy for bioabsorbable medical implants.

Bioabsorbable metallic alloys constitute a very challenging and innovative field, mainly aimed to develop the next generation of temporary medical implants. Degradation data, biological in vitro and in vivo tests are of major importance in particular for complex alloys, in which the individual element additions could enhance material performance and add functionalities. In this study, a novel Fe-Mn-Si-Cu alloy was carefully designed for vascular and blood-contact applications, and its microstructure, mechanical behavior, degradation behavior and biological performances were investigated accordingly. In previous studies, Mn and Si were found to be suitable elements to effectively enhance mechanical properties and accelerate corrosion rate in simulated body fluid. Cu was added for further grain refinement by the formation of small Cu-rich particles, potentially impacting mechanical properties and degradation behavior. In addition, the feasibility of inducing antibacterial effects in a Fe-Mn-Si-Cu alloy with low Cu content was investigated. The alloy was prepared firstly on a small scale by vacuum arc remelting, then on a larger scale by vacuum induction melting and converted into sheets by conventional thermomechanical processing techniques. Heat treatments were explored to find optimal microstructure conditions. The results confirm promising mechanical, degradation and biological performance in testing the material in in vitro conditions, showing that the degradation products are neither systematically cytotoxic nor have any hemotoxic effects. On the other hand, the expected antibacterial effects could not be confirmed.

VIEWER: an extensible visual analytics framework for enhancing mental healthcare.

A proof-of-concept study aimed at designing and implementing Visual & Interactive Engagement With Electronic Records (VIEWER), a versatile toolkit for visual analytics of clinical data, and systematically evaluating its effectiveness across various clinical applications while gathering feedback for iterative improvements. VIEWER is an open-source and extensible toolkit that employs natural language processing and interactive visualization techniques to facilitate the rapid design, development, and deployment of clinical information retrieval, analysis, and visualization at the point of care. Through an iterative and collaborative participatory design approach, VIEWER was designed and implemented in one of the United Kingdom's largest National Health Services mental health Trusts, where its clinical utility and effectiveness were assessed using both quantitative and qualitative methods. VIEWER provides interactive, problem-focused, and comprehensive views of longitudinal patient data (n = 409870) from a combination of structured clinical data and unstructured clinical notes. Despite a relatively short adoption period and users' initial unfamiliarity, VIEWER significantly improved performance and task completion speed compared to the standard clinical information system. More than 1000 users and partners in the hospital tested and used VIEWER, reporting high satisfaction and expressed strong interest in incorporating VIEWER into their daily practice. VIEWER provides a cost-effective enhancement to the functionalities of standard clinical information systems, with evaluation offering valuable feedback for future improvements. VIEWER was developed to improve data accessibility and representation across various aspects of healthcare delivery, including population health management and patient monitoring. The deployment of VIEWER highlights the benefits of collaborative refinement in optimizing health informatics solutions for enhanced patient care.

Text line segmentation approach combining deep learning model and traditional image processing techniques - application to transliteration of Cham manuscripts

Text line segmentation approach combining deep learning model and traditional image processing techniques - application to transliteration of Cham manuscripts

Treatment and Valorization of Waste Wind Turbines: Component Identification and Analysis.

Recycling end-of-life wind turbines poses a significant challenge due to the increasing number of turbines going out of use. After many years of operation, turbines lose their functional properties, generating a substantial amount of composite waste that requires efficient and environmentally friendly processing methods. Wind turbine blades, in particular, are a problematic component in the recycling process due to their complex material composition. They are primarily made of composites containing glass and carbon fibers embedded in polymer matrices such as epoxies and polyester resins. This study presents an innovative approach to analyzing and valorizing these composite wastes. The research methodology incorporates integrated processing and analysis techniques, including mechanical waste treatment using a novel compression milling process, instead of traditional knife mills, which reduces wear on the milling tools. Based on the differences in the structure and colors of the materials, 15 different kinds of samples named WT1-WT15 were distinguished from crushed wind turbines, enabling a detailed analysis of their physicochemical properties and the identification of the constituent components. Fourier transform infrared spectroscopy (FTIR) identified key functional groups, confirming the presence of thermoplastic polymers (PET, PE, and PP), epoxy and polyester resins, wood, and fillers such as glass fibers. Thermogravimetric analysis (TGA) provided insights into thermal stability, degradation behavior, and the heterogeneity of the samples, indicating a mix of organic and inorganic constituents. Differential scanning calorimetry (DSC) further characterized phase transitions in polymers, revealing variations in thermal properties among samples. The fractionation process was carried out using both wet and dry methods, allowing for a more effective separation of components. Based on the wet separation process, three fractions-GF1, GF2, and GF3-along with other components were obtained. For instance, in the case of the GF1 < 40 µm fraction, thermogravimetric analysis (TGA) revealed that the residual mass is as high as 89.7%, indicating a predominance of glass fibers. This result highlights the effectiveness of the proposed methods in facilitating the efficient recovery of high-value materials.

Psychological and Behavioral Insights From Social Media Users: Natural Language Processing-Based Quantitative Study on Mental Well-Being.

Depression significantly impacts an individual's thoughts, emotions, behaviors, and moods; this prevalent mental health condition affects millions globally. Traditional approaches to detecting and treating depression rely on questionnaires and personal interviews, which can be time consuming and potentially inefficient. As social media has permanently shifted the pattern of our daily communications, social media postings can offer new perspectives in understanding mental illness in individuals because they provide an unbiased exploration of their language use and behavioral patterns. This study aimed to develop and evaluate a methodological language framework that integrates psychological patterns, contextual information, and social interactions using natural language processing and machine learning techniques. The goal was to enhance intelligent decision-making for detecting depression at the user level. We extracted language patterns via natural language processing approaches that facilitate understanding contextual and psychological factors, such as affective patterns and personality traits linked with depression. Then, we extracted social interaction influence features. The resultant social interaction influence that users have within their online social group is derived based on users' emotions, psychological states, and context of communication extracted from status updates and the social network structure. We empirically evaluated the effectiveness of our framework by applying machine learning models to detect depression, reporting accuracy, recall, precision, and F1-score using social media status updates from 1047 users along with their associated depression diagnosis questionnaire scores. These datasets also include user postings, network connections, and personality responses. The proposed framework demonstrates accurate and effective detection of depression, improving performance compared to traditional baselines with an average improvement of 6% in accuracy and 10% in F1-score. It also shows competitive performance relative to state-of-the-art models. The inclusion of social interaction features demonstrates strong performance. By using all influence features (affective influence features, contextual influence features, and personality influence features), the model achieved an accuracy of 77% and a precision of 80%. Using affective features and affective influence features also showed strong performance, achieving 81% precision and an F1-score of 79%. The developed framework offers practical applications, such as accelerating hospital diagnoses, improving prediction accuracy, facilitating timely referrals, and providing actionable insights for early interventions in mental health treatment plans.

Speak and you shall predict: evidence that speech at initial cocaine abstinence is a biomarker of long-term drug use behavior.

Valid scalable biomarkers for predicting longitudinal clinical outcomes in psychiatric research are crucial for optimizing intervention and prevention efforts. Here we recorded spontaneous speech from initially abstinent individuals with cocaine use disorder (iCUD) for use in predicting drug use outcomes. At baseline, 88 iCUD provided 5-minute speech samples describing the positive consequences of quitting drug use and negative consequences of using drugs. Outcomes, including withdrawal, craving, abstinence days, and recent cocaine use, were assessed at three-month intervals up to one year (57 iCUD included in analyses). Predictive modeling compared natural language processing (NLP) techniques, specifically sentence embeddings with established inventories as targets, with models utilizing standard demographic and baseline psychometric variables. At short time intervals, maximal predictive power was obtained with non-NLP models that also incorporated the same drug use measures (as the outcomes) obtained at baseline, potentially reflecting their slow rate of change, which could be estimated by linear functions. However, for longer-term predictions, speech samples alone demonstrated statistically significant results, with Spearman r ≥ 0.46 and 80% accuracy for predicting abstinence. Hence speech samples may capture non-linear dynamics over extended intervals more effectively than traditional measures. These results need to be replicated in larger and independent samples. Compared to the common outcome measures used in clinical trials, speech-based measures could be leveraged as better predictors of longitudinal drug use outcomes in initially abstinent iCUD, as potentially generalizable to other subgroups with cocaine addiction, and to additional substance use disorders and related comorbidity.

Bioelectronics hydrogels for implantable cardiac and brain disease medical treatment application.

Hydrogel-based bioelectronic systems offer significant benefits for point-of-care diagnosis, treatment of cardiac and cerebral disease, surgical procedures, and other medical applications, ushering in a new era of advancements in medical technology. Progress in hydrogel-based bioelectronics has advanced from basic instrument and sensing capabilities to sophisticated multimodal perceptions and feedback systems. Addressing challenges related to immune responses and inflammation regulation after implantation, physiological dynamic mechanism, biological toxicology as well as device size, power consumption, stability, and signal conversion is crucial for the practical implementation of hydrogel-based bioelectronics in medical implants. Therefore, further exploration of hydrogel-based bioelectronics is imperative, and a comprehensive review is necessary to steer the development of these technologies for use in implantable therapies for cardiac and brain/neural conditions. In this review, a concise overview is provided on the fundamental principles underlying ionic electronic and ionic bioelectronic mechanisms. Additionally, a comprehensive examination is conducted on various bioelectronic materials integrated within hydrogels for applications in implantable medical treatments. The analysis encompasses a detailed discussion on the representative structures and physical attributes of hydrogels. This includes an exploration of their intrinsic properties such as mechanical strength, dynamic capabilities, shape-memory features, stability, stretchability, and water retention characteristics. Moreover, the discussion extends to properties related to interactions with tissues or the environment, such as adhesiveness, responsiveness, and degradability. The intricate relationships between the structure and properties of hydrogels are thoroughly examined, along with an elucidation of how these properties influence their applications in implantable medical treatments. The review also delves into the processing techniques and characterization methods employed for hydrogels. Furthermore, recent breakthroughs in the applications of hydrogels are logically explored, covering aspects such as materials, structure, properties, functions, fabrication procedures, and hybridization with other materials. Finally, the review concludes by outlining the future prospects and challenges associated with hydrogels-based bioelectronics systems.

Hate Speech Automatic Detection

This study offers a system studying-primarily based absolutely device for automated hate speech detection in textual content the usage of natural Language Processing (NLP) techniques. fashions like Logistic Regression and deep mastering algorithms (LSTM, BERT) are educated on classified datasets with features extracted via TF-IDF. The system evaluates overall overall performance using metrics together with accuracy, precision, keep in mind, and F1-score. A multilingual translation issue powered with the aid of Google Translator enables detection across languages. A Tkinter-based totally GUI allows clients to categorise remarks in actual-time, making sure accessibility and simplicity of use. elegance imbalances are addressed with weighted Logistic Regression, improving detection accuracy. The gadget demonstrates strong overall performance, promoting greater communities with the resource of offering an inexperienced device for computerized hate speech moderation.

Boosting Thermoelectric Performance of Semicrystalline Conducting Polymers by Simply Adding Nucleating Agent.

Controlling the microstructure of semiconducting polymers is critical for optimizing thermoelectric performance, yet remains challenging, requiring complex processing techniques like alignment. In this study, a straightforward strategy is introduced to enhance the thermoelectric properties of semi-crystalline polymer films by incorporating minimal amounts of nucleating agents, a method widely used in traditional polymer industries. By blending less than 1 wt% of N,N'-(1,4-phenyl)diisonicotinamide (PDA) into poly(2,5-bis(3-alkylthiophen-2-yl)thieno[3,2-b]thiophene) (PBTTT-C14), controlled modulation of crystallization behavior is achieved, resulting in reduced structural disorder and enhanced charge carrier mobility. Systematic investigations reveal that an optimal PDA loading of 0.9 wt% increases the crystallization degree by 45% compared to pristine PBTTT-C14 films. Under these optimized conditions, the PDA-modified PBTTT-C14 films exhibit a maximum electrical conductivity of 1,894Scm-1 and a maximum power factor of 176µWm-1K-2, showing improvements of 96% and 433%, respectively, over doped pristine PBTTT-C14 films. These gains are attributed to the synergistic effects of polymer chain extension and reduced grain boundary resistance, which collectively enhance charge transport efficiency. Additionally, ion exchange doping is found to maintain a high charge carrier concentration while preserving the crystallinity introduced by PDA, paving the way for advanced thermoelectric materials and next-generation polymer-based electronics.

Noise-agnostic quantum error mitigation with data augmented neural models

Quantum error mitigation, a data processing technique for recovering the statistics of target processes from their noisy version, is a crucial task for near-term quantum technologies. Most existing methods require prior knowledge of the noise model or the noise parameters. Deep neural networks have the potential to lift this requirement, but current models require training data produced by ideal processes in the absence of noise. Here we build a neural model that achieves quantum error mitigation without any prior knowledge of the noise and without training on noise-free data. To achieve this feature, we introduce a quantum augmentation technique for error mitigation. Our approach applies to quantum circuits and to the dynamics of many-body and continuous-variable quantum systems, accommodating various types of noise models. We demonstrate its effectiveness by testing it both on simulated noisy circuits and on real quantum hardware.

Pomegranate Juices: Analytical and Bio-Toxicological Comparison of Pasteurization and High-Pressure Processing in the Development of Healthy Products.

Two different produced and packaged commercial typologies of pomegranate juice were analyzed for their physicochemical, nutritional, and biological properties. The effects of classical pasteurization (PJ) and high-pressure processing (HP), applied during the productive cycle, were evaluated through several advanced analytical methods, such as CIEL*a*b* colorimetry, HPLC-DAD, DI-ESI-MS and MS/MS, and NMR analyses. Moreover, the exerted biological activity of the two pomegranate juices was monitored through Total Phenolic and Total Flavonoid Contents, antiradical, antioxidant and chelating activity. The potential inhibition of key enzymes of degenerative processes (cholinesterases, tyrosinase) and diabetes (amylase, glucosidase), the allelopathy toward Cichorium intybus, Dicondra repens, and Diplotaxis tenuifolia, and the in vivo toxicity on brine shrimp were also evaluated. The two different applied processing techniques analyzed impacted the bioactive compound's preservation differently, modifying the phytocomplex profile. HP significantly degrades punicalins and punicalagins, better preserving anthocyanins, if compared to PJ's impact. Sensory qualities, antioxidant activity, enzymatic inhibition, and ecotoxicological potential were differently impacted by the two applied processes. The obtained results can be beneficial for finding the optimal processing conditions that balance microbial safety with nutritional value preservation, contributing to the development of healthy pomegranate juice products.

Wireless power transfer empowered by reconfigurable intelligent surfaces

ABSTRACT Reconfigurable Intelligent Surfaces (RIS) represent a significant advancement in hardware technology, enhancing both wireless energy and spectral efficiency within wireless communication systems. Comprising a programmable metasurface, RIS can adeptly manipulate the wavefronts of incident signals – encompassing parameters such as phase, amplitude, frequency, and polarisation – without necessitating complex signal processing techniques. The integration of RIS into wireless communication networks enables the strategic manipulation of radio waves, thereby mitigating the adverse effects associated with natural wireless propagation. This study examines a multi-user (MU) multiple-input multiple-output (MIMO) wireless power transfer (WPT) system augmented by multiple RISs, wherein the transmitter is equipped with a constant-envelope beamformer. The optimisation of the phase shifts of the RIS and the beamformer of the transmitter is conducted jointly to maximise the total power received by users. An alternating optimisation approach is proposed, utilising the semidefinite relaxation (SDR) method. Additionally, a problem is formulated to maximise the total received power while adhering to constraints on the minimum power received by users, thereby addressing issues of user fairness. To resolve this problem, an iterative algorithm based on the Alternating Direction Method of Multipliers (ADMM) is introduced. Analysis and simulation results indicate a significant enhancement in total received power when employing the proposed methodology compared to existing algorithms documented in the literature. Furthermore, in scenarios involving multiple users, a detailed analysis of the system’s transmission performance, facilitated by RIS, is conducted. The numerical results substantiate the validity of the analysis and demonstrate the efficacy of the proposed algorithms.

BRCA Mutation Testing in Men with Metastatic Castration-Resistant Prostate Cancer: Practical Guidance for Australian Clinical Practice.

Some patients with metastatic castration-resistant prostate cancer (mCRPC) possess germline or acquired defects in the DNA damage repair (DDR) genes BRCA1 and BRCA2. Tumors with BRCA mutations exhibit sensitivity to poly-ADP ribose polymerase inhibitors (PARPi) such as olaparib and rucaparib. As a result, molecular diagnostic testing to identify patients with BRCA mutations eligible for the PARPi therapy has become an integral component of managing patients with mCRPC. There are practical challenges in the current molecular testing pathway in Australia that can compromise testing success. Testing success is often contingent on quality of tissue handling and laboratory processing techniques to minimize DNA degradation and suboptimal sequencing data quality. Greater adoption of best testing practices in Australia can be facilitated with education and greater awareness of expert recommendations. Here, we provide expert recommendations on how to optimize BRCA molecular diagnostic testing in patients with mCRPC. Optimization and standardization of molecular diagnostic testing will support health care providers and institutes in establishing more efficient testing pathways, enabling access to targeted therapies such as PARPi, and improving patient outcomes.

Exploiting latent microbial potentials for producing polyhydroxyalkanoates: A holistic approach.

Plastics are versatile, however, nonbiodegradable polymers that are primarily derived from fossil fuels and pose notable environmental challenges. However, biopolymers, such as polyhydroxyalkanoates (PHAs), poly(lactic acid), starch, and cellulose have emerged as sustainable alternatives to conventional plastics. Among these, PHAs stand out as strong contenders as they are completely bio-based and biodegradable and are synthesized by microbes as an energy reserve under stress conditions. Despite their limitations, including low mechanical strength, susceptibility to degradation, a restricted scope of application, and high production costs, biopolymers have promising potential. This review explores strategies for enhancing PHA production to address these challenges, emphasizing the need for sustainable PHA production. These strategies include selecting robust microbial strains and feedstock combinations, optimizing cell biomass and biopolymer yields, genetically engineering biosynthetic pathways, and improving downstream processing techniques. Additives such as plasticizers, thermal stabilizers, and antioxidants are crucial for modifying PHA characteristics, and its processing for achieving the desired balance between processability and end-use performance. By overcoming these complications, biopolymers have become more viable, versatile, and eco-friendly alternatives to conventional plastics, offering hope for a more sustainable future.

Predicting Risks in Healthcare Claims Using Advanced Data Processing and Machine Learning Techniques

Predicting Risks in Healthcare Claims Using Advanced Data Processing and Machine Learning Techniques

Explainable Predictive Model for Suicidal Ideation During COVID-19: Social Media Discourse Study.

Studying the impact of COVID-19 on mental health is both compelling and imperative for the health care system's preparedness development. Discovering how pandemic conditions and governmental strategies and measures have impacted mental health is a challenging task. Mental health issues, such as depression and suicidal tendency, are traditionally explored through psychological battery tests and clinical procedures. To address the stigma associated with mental illness, social media is used to examine language patterns in posts related to suicide. This strategy enhances the comprehension and interpretation of suicidal ideation. Despite easy expression via social media, suicidal thoughts remain sensitive and complex to comprehend and detect. Suicidal ideation captures the new suicidal statements used during the COVID-19 pandemic that represents a different context of expressions. In this study, our aim was to detect suicidal ideation by mining textual content extracted from social media by leveraging state-of-the-art natural language processing (NLP) techniques. The work was divided into 2 major phases, one to classify suicidal ideation posts and the other to extract factors that cause suicidal ideation. We proposed a hybrid deep learning-based neural network approach (Bidirectional Encoder Representations from Transformers [BERT]+convolutional neural network [CNN]+long short-term memory [LSTM]) to classify suicidal and nonsuicidal posts. Two state-of-the-art deep learning approaches (CNN and LSTM) were combined based on features (terms) selected from term frequency-inverse document frequency (TF-IDF), Word2vec, and BERT. Explainable artificial intelligence (XAI) was used to extract key factors that contribute to suicidal ideation in order to provide a reliable and sustainable solution. Of 348,110 records, 3154 (0.9%) were selected, resulting in 1338 (42.4%) suicidal and 1816 (57.6%) nonsuicidal instances. The CNN+LSTM+BERT model achieved superior performance, with a precision of 94%, a recall of 95%, an F1-score of 94%, and an accuracy of 93.65%. Considering the dynamic nature of suicidal behavior posts, we proposed a fused architecture that captures both localized and generalized contextual information that is important for understanding the language patterns and predict the evolution of suicidal ideation over time. According to Local Interpretable Model-Agnostic Explanations (LIME) and Shapley Additive Explanations (SHAP) XAI algorithms, there was a drift in the features during and before COVID-19. Due to the COVID-19 pandemic, new features have been added, which leads to suicidal tendencies. In the future, strategies need to be developed to combat this deadly disease.