17628 research outputs found
Sort by
A global assessment of BirdNET performance: Differences among continents, biomes, and species
Full text linkRecent advances in machine learning have accelerated automated species detection across diverse ecological domains, enabling large-scale, non-invasive monitoring of biodiversity. In ornithological research, the combination of passive acoustic monitoring (PAM) and rapidly-developing novel identification tools such as BirdNET—a deep learning–based sound recognition algorithm—offers new opportunities for surveying vocally active bird communities. Here, we present the first worldwide evaluation of BirdNET using 4224 one-minute recordings from 67 sites across all continents annotated by local experts. More specifically, we assessed the capacity of BirdNET to accurately identify individual vocalizations and characterize bird communities based on the automated analysis of passively collected soundscapes. We further analyzed how its performance varies across continents, biomes, species, and minimum confidence thresholds. The proportion of correct BirdNET predictions (precision) was generally high and consistent across continents (range: 0.57–0.71) and biomes (range: 0.55–0.76). In contrast, the proportion of vocalizations successfully detected (recall) was generally lower and more heterogeneous across continents (range: 0.24–0.52) and biomes (range: 0.34–0.72), reflecting differences in species coverage and local ecological context. BirdNET predictive power, as measured by the Precision-Recall Area Under the Curve (PR AUC; higher values indicating better performance), was highest in North America, Oceania, and Europe (range: 0.16–0.23), moderate in Central/South America (0.13), and lowest in Africa and Asia (range: 0.03–0.04). Species-specific analyses revealed substantial heterogeneity in detection accuracy, with optimal confidence thresholds varying widely by species and analytical goal. Our results establish a global reference point for BirdNET reliability and highlight where algorithmic refinement and expanded acoustic sampling are most needed
Hybrid Poly Commitments for Scalable Binius Zero-Knowledge Proofs in Federated Learning
No full textFederated learning enables collaborative model training without sharing raw data, but practical deployments increasingly require verifiable guarantees that clients compute updates correctly. Zero-knowledge proofs can provide such guarantees, yet existing approaches face scalability limits due to the combined cost of polynomial commitments and fast Fourier transform (FFT) intensive verification. Pairing-based schemes offer compact proofs but incur high prover and verifier overhead, while hash-based constructions reduce algebraic cost at the expense of rapidly growing proof sizes. This paper proposes Hybrid-Commit, a polynomial commitment architecture for Binius zero-knowledge proofs that aligns cryptographic primitives with the algebraic structure of federated learning workloads. The scheme separates verification into additive and multiplicative phases: linear aggregation is handled using batched additive commitments optimized for binary fields, while non-linear constraints are verified via hash-based commitments over sparsely selected FFT domains. Proofs from multiple clients are combined through recursive aggregation while preserving non-interactivity. Experiments demonstrate scalability in prover time and proof size (near-constant prover time across 4–11 clients; 160 bytes per client representing 341× and 813× reductions vs. FRI-PCS and Orion), although verification time (762 ms per client) does not scale favorably, making the scheme suitable for bandwidth-constrained scenarios. The scheme achieves under 2% end-to-end training overhead with no impact on model accuracy, indicating that workload-aware commitment design can improve specific scalability dimensions of zero-knowledge verification in federated learning systems
The relationship between socially aversive personality of project managers and project performance: Evidence from the UAE
Full text linkWe study the relationship between socially aversive personality of project managers that influence behaviour, and project performance. Data from 409 project managers in the United Arab Emirates were collected using a 36-factor, workplace-focused dark personality questionnaire adapted from existing validated scales. Analysis was via Structural Equation Modelling (SPSS AMOS 29). Findings suggest that project managers: (i) exhibited higher Narcissism levels than Psychopathy or Machiavellianism; (ii) showed no significant variation in dark trait subscales by project characteristics; and (iii) display positively correlated dark trait subscales. Results also indicate: (iv) higher levels of each dark trait corresponding with a relationship to poorer project performance; and (v) Psychopathy's independent effect not being statistically significant. Given their negative relationship with project outcomes, organisations should prioritize identifying and managing these traits. Given the paucity of studies relating to this research, our study was exploratory in nature. Thus, this study serves as a primer on project-focused dark personality research and offers a novel perspective on the antecedents of project performance
CANNSkin: A Convolutional Autoencoder Neural Network-Based Model for Skin Cancer Classification
Full text linkVisual diagnosis of skin cancer is challenging due to subtle inter-class similarities, variations in skin texture, the presence of hair, and inconsistent illumination. Deep learning models have shown promise in assisting early detection, yet their performance is often limited by the severe class imbalance present in dermoscopic datasets. This paper proposes CANNSkin, a skin cancer classification framework that integrates a convolutional autoencoder with latent-space oversampling to address this imbalance. The autoencoder is trained to reconstruct lesion images, and its latent embeddings are used as features for classification. To enhance minority-class representation, the Synthetic Minority Oversampling Technique (SMOTE) is applied directly to the latent vectors before classifier training. The encoder and classifier are first trained independently and later fine-tuned end-to-end. On the HAM10000 dataset, CANNSkin achieves an accuracy of 93.01%, a macro-F1 of 88.54%, and an ROC–AUC of 98.44%, demonstrating strong robustness across ten test subsets. Evaluation on the more complex ISIC 2019 dataset further confirms the model’s effectiveness, where CANNSkin achieves 94.27% accuracy, 93.95% precision, 94.09% recall, and 99.02% F1-score, supported by high reconstruction fidelity (PSNR 35.03 dB, SSIM 0.86). These results demonstrate the effectiveness of our proposed latent-space balancing and fine-tuned representation learning as a new benchmark method for robust and accurate skin cancer classification across heterogeneous datasets
Clinician assessed rates of PTSD and Complex PTSD in a medical-rehabilitation sample of active-duty military personnel in the Armed Forces of Ukraine
No full textIntroduction: Millions of people have served in the Armed Forces of Ukraine (AFU) since Russia's invasion in 2014, but there is currently no information about the prevalence of posttraumatic stress disorder (PTSD) in this population. The main purpose of this study was to estimate rates of ICD‐11 PTSD and Complex PTSD (CPTSD), and comorbidity with major depression, in a sample of active‐duty, combat‐exposed AFU military personnel. Methods: Clinical interviews were conducted with 590 soldiers recruited from military hospitals and rehabilitation centers in Ukraine. All were trauma‐exposed during military operations. PTSD and CPTSD were diagnosed using the International Trauma Interview, and a current episode of major depression was diagnosed using the Mini‐International Neuropsychiatric Interview. Results: Overall, 67.4% of soldiers were diagnosed with ICD‐11 PTSD or CPTSD, with 45.9% being diagnosed with PTSD and 21.5% with CPTSD. Additionally, 34.4% were diagnosed with major depression, and comorbidity with PTSD (45.0%) and CPTSD (51.2%) was high. Elevated rates of PTSD were observed for current smokers and those who were currently consuming alcohol, while elevated rates of CPTSD were observed for officers (versus enlisted soldiers) and those recruited from rehabilitation facilities (vs. general hospitals). Conclusion: Although not representative of the entire AFU population, these results imply that hundreds of thousands of soldiers (and veterans) in Ukraine are likely experiencing clinically significant posttraumatic distress related to their combat experiences. Results are discussed in the context of finding scalable approaches to addressing this mental health challenge
Translation and validation of the Czech Partner version of the Birth Satisfaction Scale-Revised (BSS-R)
Full text linkBackgroundThe Birth Satisfaction Scale-Revised (BSS-R) is a widely used, psychometrically robust and brief self-report measure of birth experience from the mothers perspective. The current study sought to adapt and validate the BSS-R for partners, evaluating key psychometric properties, including the underlying tri-dimensional factor structure of stress experienced, personal attributes and quality of care.AimTo translate and validate a Czech speaking partner version of the Birth Satisfaction Scale-Revised (BSS-R) and examine key measurement characteristics and association with fundamental clinical outcome variables.MethodFollowing translation of the UK partner BSS-R into Czech, the Czech Partner BSS-R (CZP-BSS-R) was administered to 225 partners of women who had given birth within the past 5-years. Key psychometric characteristics were examined, including factor structure, divergent and known-groups discriminant validity and internal reliability,ResultsEstablished measurement models of the BSS-R observed in mothers were found to offer an excellent fit to partner data. The CZP-BSS-R also demonstrated excellent validity and reliability characteristics.ConclusionsThe CZP-BSS-R was found to be valid and reliable, with results from Czech partners ‘mirroring’ factor structure and key validity characteristics previously established in Czech mothers. The BSS-R validated for completion by Czech speaking mothers now has a matched version available for use with Czech speaking partners
Talent identification and development strategies in elite women’s soccer: a pan-European perspective
Full text linkThe question of how best to identify and develop youth soccer players has received considerable attention from the scientific community. Existing literature has, however, largely focused on male players, with comparatively little exploration of the specific approaches employed within women’s soccer. Accordingly, we sought to investigate the key factors deemed important by elite women’s soccer clubs concerning the: 1) identification of potential talent; 2) development of players within the player pathway; and 3) selection of players for the next age group or senior team. Data were generated through semi-structured interviews with 11 key representatives from seven elite women’s soccer clubs. Clubs were purposefully sampled to include the highest performing teams (38 domestic titles and 10 UEFA Women’s Champions League titles) from five European nations (Spain, France, Sweden, Germany, and Italy). Data were analysed using thematic content analysis, resulting in six higher-order themes: 1) prioritising local talent; 2) recruitment from mixed grassroots leagues; 3) creating challenging developmental environments; 4) ensuring player wellbeing; 5) patience in decision-making; and 6) facilitating the youth-to-senior transition through a top-down approach. A total of 17 lower-order themes were subsequently identified. The present study offers novel insights of key strategies deemed important by some of the most successful women’s clubs in top-performing European nations. Future research examining the efficacy of such approaches could help inform the development of evidence-based practices for nurturing the next generation of elite female players
Generative Adversarial Networks-enabled Anomaly Detection Systems: A Survey
No full textAnomaly Detection (AD) is an important area of research because it helps identify outliers in data, enabling early detection of errors, fraud, and potential security breaches. Machine Learning (ML) can be utilized for distinct AD systems, and Generative Adversarial Networks (GANs) have emerged as a promising technique due to their ability to generate new data that closely resembles a given dataset, allowing for the creation of realistic images, videos, audio, text, and other types of synthetic data. This paper explores state-of-the-art approaches in AD using GANs. The paper starts by providing a comprehensive overview of ML techniques for AD, including supervised, unsupervised, and semi-supervised approaches. This survey also explores various AD approaches based on GANs and provides an application-based classification of GANs-based AD approaches in the Internet-of-Things (IoT), Industrial IoT, Digital Healthcare, Energy Management Systems, and Cellular Network domains. Moreover, the paper discusses several datasets used in evaluating the performance of GANs-based AD techniques such as BOT-IoT, TON-IoT, CIC-IoT, CIC-IDS, and NSL-KDD. These datasets serve as valuable resources for researchers and practitioners to develop and test AD systems, particularly in the context of IoT and network security. Furthermore, the paper discusses the challenges and limitations of GANs-based AD techniques and proposes future research directions to address these challenges
A comprehensive review of frequency response and control strategies for grid-connected solar photovoltaic systems
Full text linkRenewable energy is the future of energy generation to meet the growing electricity demand worldwide. The extensive integration of Renewable Energy Sources (RESs) is causing a paradigm shift in the power network. Integrating RESs reduces the overall inertia of the system, which could result in occasional unstable frequency and may lead to cascading blackouts. This paper performs an overarching analysis of different frequency control techniques that support seamless integration of solar photovoltaic systems to the grid. These techniques are extensively compared based on various parameters and are categorised based on control actions and implementation to highlight their effectiveness and limitations. Before exploring frequency regulation methods, a primer on PV system modelling and configuration is given to comprehend PV integration to the grid. Furthermore, a brief synopsis of strategies utilising Energy Storage Devices (ESSs) is provided, emphasising methods that do not require ESSs. Beyond this, an overview of optimisation algorithms used in the literature for frequency regulation is given, providing a holistic understanding. Recent research highlights the growing interest in hybrid power reserve approaches, as combining diverse optimisation algorithms offers robust, reliable, fast, and adaptive solutions. This paper endeavours to provide a holistic review for researchers interested in developing frequency regulation methods for PV systems and to support industry practitioners in finding the appropriate solutions for frequency regulation-related applications
Exploring human milk donation: A cross-sectional study
Full text linkObjective To examine and describe the current practices and perspectives regarding human milk donation in Belgium.Study design A cross-sectional study was conducted, utilising an online survey distributed to women and their partners or co-parents.Methods Demographic and personal details and responses related to awareness, information resources, milk transfer, and donor-recipient and recipient-donor contact were obtained. We measured factors influencing the intention to donate and receive human milk (0–5 scale), attitudes toward human milk donation (1–4 scale) and donation practices (dichotomous responses). Descriptive statistics were used to analyse the data.Results The analysis included 873 respondents. A surplus (4.6 ± 1.0) or shortage (3.9 ± 1.7) of milk is the primary factor influencing the intention to donate or receive human milk, reported by 88 % and 78 % of respondents, respectively. Most donations (85.7 %) occur through social media (50 %) and personal networks (47.6 %). Key perspectives of milk donation include healthcare providers’ familiarity (3.7 ± 0.5), public awareness (3.6 ± 0.5), mandatory blood testing (3.5 ± 0.7), and donor screening (3.5 ± 0.6). Human milk banks are perceived as valuable resources (3.5 ± 0.6), while altruism (3.4 ± 0.6) is key to donating. There is broad support for making human milk accessible to all infants, not just those who are premature or ill (3.4 ± 0.6).Conclusions Personal attitudes and intentions shape informal, self-regulated milk donation. Healthcare providers should consider these perspectives when advising on milk donation practices
Data sources:
Access Repository Dashboard
Do you manage Repository@Napier? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!