Search Results (59)

Credit rating transitions and defaults are critical indicators of corporate creditworthiness, yet their accurate modeling remains a persistent challenge in risk management. Traditional models such as logistic regression (LR) and structural approaches (e.g., Merton’s model) offer transparency but often fail to capture nonlinear relationships, temporal dynamics, and firm heterogeneity. This study proposes a hybrid machine learning (ML) framework to explain and predict corporate rating transitions and defaults, addressing key limitations in existing literature. We benchmark four classification algorithms—LR, Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Support Vector Machines (SVM)—on a structured corporate credit dataset. Our approach integrates segment-specific modeling across rating bands, out-of-time validation to simulate real-world applicability, and SHapley Additive exPlanations (SHAP) values to ensure interpretability. The results demonstrate that ensemble methods, particularly XGBoost and RF, significantly outperform LR and SVM in predictive accuracy and early warning capability. Moreover, SHAP analysis reveals differentiated drivers of rating transitions across credit quality segments, highlighting the importance of tailored monitoring strategies. This research contributes to the literature by bridging predictive performance with interpretability in credit risk modeling and offers practical implications for regulators, rating agencies, and financial institutions seeking robust, transparent, and forward-looking credit assessment tools. Full article

This study investigates the nonlinear and regime-dependent relationship between climate policy uncertainty (CPU) and sovereign credit default swap (CDS) spreads across a panel of developed and emerging economies from February 2010 to March 2025. Utilizing the Quantile-on-Quantile Regression (QQR) and Multivariate QQR (MQQR) frameworks, we capture the heterogeneous effects of CPU under varying market states and assess the marginal role of global risk factors, including geopolitical risk (GPR), economic policy uncertainty (EPU), and market volatility (VIX). The findings indicate that in developed markets, CPU exerts a nonlinear impact that intensifies during periods of heightened sovereign risk, while in low-risk regimes, its effect is often muted or reversed. In contrast, emerging economies exhibit more volatile and state-contingent responses, with CPU exerting stronger effects in calm conditions but diminishing in explanatory power once global risks are taken into account. These dynamics highlight the importance of institutional credibility and financial integration in moderating CPU-driven credit risk. Full article

Borderline personality disorder (BPD) is a complex and heterogeneous condition characterized by emotional instability, impulsivity, and impaired regulation of interpersonal relationships. This narrative review integrates findings from recent neuroimaging, neurochemical, and treatment studies to identify core neurobiological mechanisms and highlight translational potential. Evidence from 112 studies published up to 2025 is synthesized, encompassing structural MRI, resting-state and task-based functional MRI, EEG, PET, and emerging machine learning applications. Consistent disruptions are observed across the prefrontal–amygdala circuitry, the default mode network (DMN), and mentalization-related regions. BPD shows a dominant and stable pattern of hyperconnectivity in the precuneus. Transdiagnostic comparisons with PTSD and cocaine use disorder (CUD) suggest partial overlap in DMN dysregulation, though BPD-specific traits emerge in network topology. Machine learning models achieve a classification accuracy of 70–88% and may support the tracking of early treatment responses. Longitudinal fMRI studies indicate that psychodynamic therapy facilitates the progressive normalization of dorsal anterior cingulate cortex (dACC) activity and reductions in alexithymia. We discuss the role of phenotypic heterogeneity (internalizing versus externalizing profiles), the potential of neuromodulation guided by biomarkers, and the need for standardized imaging protocols. Limitations include small sample sizes, a lack of effective connectivity analyses, and minimal multicenter cohort representation. Future research should focus on constructing multimodal biomarker panels that integrate functional connectivity, epigenetics, and computational phenotyping. This review supports the use of a precision psychiatry approach for BPD by aligning neuroscience with scalable clinical tools. Full article

Introduction: Substance use disorders (SUDs) pose a significant public health challenge, with current treatments often exhibiting limited effectiveness and high relapse rates. Transcranial direct current stimulation (tDCS), a noninvasive neuromodulation technique that delivers low-intensity direct current via scalp electrodes, has shown promise in various psychiatric and neurological conditions. In SUDs, tDCS may help to modulate key neurocircuits involved in craving, executive control, and reward processing, potentially mitigating compulsive drug use. However, the precise neurobiological mechanisms by which tDCS exerts its therapeutic effects in SUDs remain only partly understood. This review addresses that gap by synthesizing evidence from clinical studies that used neuroimaging (fMRI, fNIRS, EEG) and blood-based biomarkers to elucidate tDCS’s mechanisms in treating SUDs. Methods: A targeted literature search identified articles published between 2008 and 2024 investigating tDCS interventions in alcohol, nicotine, opioid, and stimulant use disorders, focusing specifically on physiological and neurobiological assessments rather than purely behavioral outcomes. Studies were included if they employed either neuroimaging (fMRI, fNIRS, EEG) or blood tests (neurotrophic and neuroinflammatory markers) to investigate changes induced by single- or multi-session tDCS. Two reviewers screened titles/abstracts, conducted full-text assessments, and extracted key data on participant characteristics, tDCS protocols, neurobiological measures, and clinical outcomes. Results: Twenty-seven studies met the inclusion criteria. Across fMRI studies, tDCS—especially targeting the dorsolateral prefrontal cortex—consistently modulated large-scale network activity and connectivity in the default mode, salience, and executive control networks. Many of these changes correlated with subjective craving, attentional bias, or extended time to relapse. EEG-based investigations found that tDCS can alter event-related potentials (e.g., P3, N2, LPP) linked to inhibitory control and salience processing, often preceding or accompanying changes in craving. One fNIRS study revealed enhanced connectivity in prefrontal regions under active tDCS. At the same time, two blood-based investigations reported the partial normalization of neurotrophic (BDNF) and proinflammatory markers (TNF-α, IL-6) in participants receiving tDCS. Multi-session protocols were more apt to drive clinically meaningful neuroplastic changes than single-session interventions. Conclusions: Although significant questions remain regarding optimal stimulation parameters, sample heterogeneity, and the translation of acute neural shifts into lasting behavioral benefits, this research confirms that tDCS can induce detectable neurobiological effects in SUD populations. By reshaping activity across prefrontal and reward-related circuits, modulating electrophysiological indices, and altering relevant biomarkers, tDCS holds promise as a viable, mechanism-based adjunctive therapy for SUDs. Rigorous, large-scale studies with longer follow-up durations and attention to individual differences will be essential to establish how best to harness these neuromodulatory effects for durable clinical outcomes. Full article

Climate change risk has significant impacts on corporate financial activities. Using firm-level data from A-share listed companies in China from 2010 to 2022, we examine how climate risk affects corporate debt financing capacity. We find that climate change risk significantly weakens firms’ ability to raise debt, leading to lower leverage and higher financing costs. These results remain robust across various checks for endogeneity and alternative specifications. We also show that reducing corporate carbon emission intensity can mitigate the negative impact of climate risk on debt financing, suggesting that supply-side credit policies are more effective than demand-side capital structure choices. Furthermore, we identify three channels through which climate risk impairs debt capacity: reduced competitiveness, increased default risk, and diminished resilience. Our heterogeneity analysis reveals that these adverse effects are more pronounced for non-state-owned firms, firms with weaker internal controls, and companies in highly financialized regions, and during periods of heightened environmental uncertainty. We also apply textual analysis and machine learning to the measurement of climate change risks, partially mitigating the geographic biases and single-dimensional shortcomings inherent in macro-level indicators, thus enriching the quantitative research on climate change risks. These findings provide valuable insights for policymakers and financial institutions in promoting corporate green transition, guiding capital allocation, and supporting sustainable development. Full article

Intracranial tumors such as gliomas, meningiomas, and brain metastases induce complex alterations in brain function beyond their focal presence. Modern connectomic and neuroimaging approaches, including resting-state functional MRI (rs-fMRI) and diffusion MRI, have revealed that these tumors disrupt and reorganize large-scale brain networks in heterogeneous ways. In adult patients, diffuse gliomas infiltrate neural circuits, causing both local disconnections and widespread functional changes that often extend into structurally intact regions. Meningiomas and metastases, though typically well-circumscribed, can perturb networks via mass effect, edema, and diaschisis, sometimes provoking global “dysconnectivity” related to cognitive deficits. Therefore, this review synthesizes interdisciplinary evidence from neuroscience, oncology, and neuroimaging on how intracranial tumors disrupt functional brain connectivity pre- and post-surgery. We discuss how functional heterogeneity (i.e., differences in network involvement due to tumor type, location, and histo-molecular profile) manifests in connectomic analyses, from altered default mode and salience network activity to changes in structural–functional coupling. The clinical relevance of these network effects is examined, highlighting implications for pre-surgical planning, prognostication of neurocognitive outcomes, and post-operative recovery. Gliomas demonstrate remarkable functional plasticity, with network remodeling that may correlate with tumor genotype (e.g., IDH mutation), while meningioma-related edema and metastasis location modulate the extent of network disturbance. Finally, we explore future directions, including imaging-guided therapies and “network-aware” neurosurgical strategies that aim to preserve and restore brain connectivity. Understanding functional heterogeneity in brain tumors through a connectomic lens not only provides insights into the neuroscience of cancer but also informs more effective, personalized approaches to neuro-oncologic care. Full article

Background: Olfactory dysfunction (OD)—including anosmia and hyposmia—is a common and often persistent outcome of viral infections. This systematic review consolidates findings from structural and functional MRI studies to explore how COVID-19 SARS-CoV-2-induced smell loss alters the brain. Considerable heterogeneity was observed across studies, influenced by differences in methodology, population characteristics, imaging timelines, and OD classification. Methods: Following PRISMA guidelines, we conducted a systematic search of PubMed/MEDLINE, Scopus, and Web of Science to identify MRI-based studies examining COVID-19’s SARS-CoV-2 OD. Twenty-four studies were included and categorized based on imaging focus: (1) olfactory bulb (OB), (2) olfactory sulcus (OS), (3) grey and white matter changes, (4) task-based brain activation, and (5) resting-state functional connectivity. Demographic and imaging data were extracted and analyzed accordingly. Results: Structural imaging revealed consistent reductions in olfactory bulb volume (OBV) and olfactory sulcus depth (OSD), especially among individuals with OD persisting beyond three months, suggestive of inflammation and neurodegeneration in olfactory-associated regions like the orbitofrontal cortex and thalamus. Functional MRI studies showed increased connectivity in early-stage OD within regions such as the piriform and orbitofrontal cortices, possibly reflecting compensatory activity. In contrast, prolonged OD was associated with reduced activation and diminished connectivity, indicating a decline in olfactory processing capacity. Disruptions in the default mode network (DMN) and limbic areas further point to secondary cognitive and emotional effects. Diffusion tensor imaging (DTI) findings—such as decreased fractional anisotropy (FA) and increased mean diffusivity (MD)—highlight white matter microstructural compromise in individuals with long-term OD. Conclusions: COVID-19’s SARS-CoV-2 olfactory dysfunction is associated with a range of cerebral alterations that evolve with the duration and severity of smell loss. Persistent dysfunction correlates with greater neural damage, underscoring the need for longitudinal neuroimaging studies to better understand recovery dynamics and guide therapeutic strategies. Full article

The adoption of Machine Learning Operations (MLOps) has grown rapidly as organisations seek to streamline the development and deployment of machine learning (ML) models. A core concept in MLOps workflows is the ML pipeline, consisting of a sequence of tasks representing the various stages of the ML lifecycle, such as data preprocessing, model training, and evaluation. As these tasks have different resource requirements and computational demands, using heterogeneous computing environments has become important. However, to exploit this heterogeneity, it is essential to map each task within a pipeline to the right machine. This paper introduces a modular and flexible placement system for ML pipelines that automatically allocates tasks to the most suitable machines in order to reduce execution and waiting times. Although designed to support custom placement strategies, the system employs a two-phase strategy: pipeline scheduling and task placement. During the scheduling phase, the Shortest Job First (SJF) algorithm determines the execution order of the pipelines. In the task placement phase, a heuristic-based method is used to assign tasks to machines. Experimental evaluations across a range of ML models and datasets demonstrate that the proposed system significantly outperforms baseline methods and the Kubernetes default scheduler. It achieved reductions of up to 68% in total execution time, and over 80% in average waiting time. Moreover, the system also demonstrates efficient pipeline dispatching in scenarios where multiple pipelines are submitted for execution. These results highlight the system’s potential to improve resource utilisation and accelerate ML model development in heterogeneous environments. Full article

Credit risk assessment plays a critical role in financial risk management, focusing on predicting borrower default to minimize losses and ensure compliance. This study systematically reviews 23 empirical articles published between 2019 and 2023, highlighting the integration of machine learning and optimization techniques, particularly bio-inspired metaheuristics, for feature selection in individual credit risk assessment. These nature-inspired algorithms, derived from biological and ecological processes, align with bio-inspired principles by mimicking natural intelligence to solve complex problems in high-dimensional feature spaces. Unlike prior reviews that adopt broader scopes combining corporate, sovereign, and individual contexts, this work focuses exclusively on methodological strategies for individual credit risk. It categorizes the use of machine learning algorithms, feature selection methods, and metaheuristic optimization techniques, including genetic algorithms, particle swarm optimization, and biogeography-based optimization. To strengthen transparency and comparability, this review also synthesizes classification performance metrics—such as accuracy, AUC, F1-score, and recall—reported across benchmark datasets. Although no unified experimental comparison was conducted due to heterogeneity in study protocols, this structured summary reveals consistent trends in algorithm effectiveness and evaluation practices. The review concludes with practical recommendations and outlines future research directions to improve fairness, scalability, and real-time application in credit risk modeling. Full article

Background/Objectives: The severity of autism spectrum disorder (ASD) is clinically assessed through a comprehensive evaluation of social communication deficits, restricted interests, repetitive behaviors, and the level of support required (ranging from level 1 to level 3) according to DSM-5 criteria. Along with its varied clinical manifestations, the neuroanatomy of ASD is characterized by heterogeneous abnormalities. Notably, brain MRI of children with ASD often reveals an increased number of perivascular spaces (PVSs) compared to typically developing children. Our recent findings indicate that enlarged PVSs (ePVSs) are more common in younger male patients with severe ASD and that specific ePVS locations are significantly associated with ASD symptoms. Methods: In this study, we mapped ePVSs across key regions of three major cognitive networks—the Default Mode Network (DMN), the combined Central Executive/Frontoparietal Network (CEN/FPN), and the Salience Network (SN)—in 36 individuals with different symptom severities and rehabilitation needs due to ASD. We explored how the number, size, and location of PVSs in these networks are related to specific ASD symptoms and the overall need for rehabilitation and support. Results: Our results suggest that ePVSs in the DMN, CEN/FPN, and SN are strongly correlated with the severity of certain ASD symptoms, including verbal deficits, stereotypies, and sensory disturbances. We found a mild association between ePVSs and the level of support needed for daily living and quality of life. Conclusions: Dysfunction in cognitive networks associated with the presence of ePVSs has a significant impact on the severity of ASD symptoms. However, the need for assistance may also be influenced by other comorbid conditions and dysfunctions in smaller, overlapping brain networks. Full article

In this paper, we tackle the challenge of optimizing resource utilization and demand-driven allocation in dynamic, multi-dimensional heterogeneous environments. Traditional containerized task scheduling systems, like Kubernetes, typically rely on default schedulers that primarily focus on CPU and memory, overlooking the multi-dimensional nature of heterogeneous resources such as GPUs, network I/O, and disk I/O. This results in suboptimal scheduling and underutilization of resources. To address this, we propose a dynamic scheduling method for heterogeneous resources using an enhanced Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) algorithm that adjusts weights in real time and applies nonlinear normalization. Leveraging parallel computing, approximation, incremental computation, local updates, and hardware acceleration, the method minimizes overhead and ensures efficiency. Experimental results showed that, under low-load conditions, our method reduced task response times by 31–36%, increased throughput by 20–50%, and boosted resource utilization by over 20% compared to both the default Kubernetes scheduler and the Kubernetes Container Scheduling Strategy (KCSS) algorithm. These improvements were tested across diverse workloads, utilizing CPU, memory, GPU, and I/O resources, in a large-scale cluster environment, demonstrating the method’s robustness. These enhancements optimize cluster performance and resource efficiency, offering valuable insights for task scheduling in containerized cloud platforms. Full article

Information asymmetry between banks and enterprises in the credit market is essentially the microfoundation of financial risk generation. The frequent occurrence of corporate debt defaults, mainly due to the behavior of short-term debt for long-term use (hereinafter referred to as “SDLU”), further aggravates the contagion path from individual liquidity crisis to systemic repayment crisis. In order to test whether bank financial technology (hereinafter referred to as “BankFintech”) can mitigate SDLU and reduce the possibility of financial risks, this study matched the loan data of China’s A-share listed companies with the patent data of bank-invented Fintech from 2013 to 2022 to construct the BankFintech Development Index for empirical analysis. The empirical results show that the development of BankFintech can significantly inhibit SDLU. The mechanism test reveals that BankFintech reduces bank credit risk and liquidity risk by lowering firms’ risk-weighted assets, improving capital adequacy and liquidity ratios, tilts banks’ lending preferences toward duration-matched long-term financing, and “forces” enterprises to take the initiative to improve their financial health and information transparency, enhance their ability to obtain long-term loans, and realize the active management of mismatch risk. Heterogeneity analysis finds that the effect is more significant in non-state-owned enterprises and technology-intensive industries. Further analysis shows that the level of enterprise digitization, the intensity of financial regulation, and related financial policies significantly moderate the marginal effect between the two. This study verified the “Porter’s Risk Mitigation Hypothesis” of Fintech, providing empirical evidence for effectively cracking the financial vulnerability caused by debt maturity mismatch and deepening financial supply-side reform. Full article

As global climate change issues have become increasingly severe, green finance has gained widespread attention from governments and financial institutions as a crucial tool for promoting sustainable development. This paper explores the impact of green finance reform pilot zones on corporate debt default risks based on a difference-in-differences model. We found that green finance policies significantly increase corporate debt default risks by exacerbating financing constraints and reducing stock liquidity. A heterogeneity analysis revealed that polluting enterprises, non-state-owned enterprises, and companies in the Eastern region are more susceptible to the impacts of this policy. This paper suggests that the government should formulate differentiated green finance policies tailored to different types of enterprises and regional characteristics. Full article

As monitoring technologies and data collection methodologies advance, landslide disaster data reflects attributes such as diverse sources, heterogeneity, substantial volumes, and stringent real-time requirements. To bolster the data support capabilities for the monitoring, prevention, and management of landslide disasters, the efficient integration of multi-source heterogeneous data is of paramount importance. The present study proposes an innovative approach to integrate multi-source landslide disaster data by combining the Flink-oriented framework with load balancing task scheduling based on an improved particle swarm optimization (APSO) algorithm. It utilizes Flink’s streaming processing capabilities to efficiently process and store multi-source landslide data. To tackle the issue of uneven cluster load distribution during the integration process, the APSO algorithm is proposed to facilitate cluster load balancing. The findings indicate the following: (1) The multi-source data integration method for landslide disaster based on Flink and APSO proposed in this article, combined with the structural characteristics of landslide disaster data, adopts different integration methods for data in different formats, which can effectively achieve the integration of multi-source landslide data. (2) A multi-source landslide data integration framework based on Flink has been established. Utilizing Kafka as a message queue, a real-time data pipeline was constructed, with Flink facilitating data processing and read/write operations for the database. This implementation achieves efficient integration of multi-source landslide data. (3) Compared to Flink’s default task scheduling strategy, the cluster load balancing strategy based on APSO demonstrated a reduction of approximately 4.7% in average task execution time and an improvement of approximately 5.4% in average system throughput during actual tests using landslide data sets. The research findings illustrate a significant improvement in the efficiency of data integration processing and system performance. Full article

Schizophrenia, a highly complex psychiatric disorder, presents significant challenges in diagnosis and treatment due to its multifaceted neurobiological underpinnings. Recent advancements in functional magnetic resonance imaging (fMRI) and artificial intelligence (AI) have revolutionized the understanding and management of this condition. This manuscript explores how the integration of these technologies has unveiled key insights into schizophrenia’s structural and functional neural anomalies. fMRI research highlights disruptions in crucial brain regions like the prefrontal cortex and hippocampus, alongside impaired connectivity within networks such as the default mode network (DMN). These alterations correlate with the cognitive deficits and emotional dysregulation characteristic of schizophrenia. AI techniques, including machine learning (ML) and deep learning (DL), have enhanced the detection and analysis of these complex patterns, surpassing traditional methods in precision. Algorithms such as support vector machines (SVMs) and Vision Transformers (ViTs) have proven particularly effective in identifying biomarkers and aiding early diagnosis. Despite these advancements, challenges such as variability in methodologies and the disorder’s heterogeneity persist, necessitating large-scale, collaborative studies for clinical translation. Moreover, ethical considerations surrounding data integrity, algorithmic transparency, and patient individuality must guide AI’s integration into psychiatry. Looking ahead, AI-augmented fMRI holds promise for tailoring personalized interventions, addressing unique neural dysfunctions, and improving therapeutic outcomes for individuals with schizophrenia. This convergence of neuroimaging and computational innovation heralds a transformative era in precision psychiatry. Full article