151398 research outputs found
Sort by
Unlocking the mysteries of drought: integrating snowmelt dynamics into drought analysis at the Narayani River Basin, Nepal
No full textDrought is among the most impactful natural hazards, undermining water security, agriculture, and livelihoods worldwide. Analysing droughts in large catchments presents several unique challenges, primarily related to the complexity of land surface characteristics and data availability limitations. Conducting drought analysis in the Narayani River Basin, which encompasses a vast area within the Himalayan region of Nepal, is extremely challenging but crucial for maintaining the river basin's social, economic, and environmental balance. In response, this study develops a new combined drought index (CDI), integrating satellite-based reanalysis parameters [i.e., Land Surface Temperature (LST), Snow Cover (SC), and Normalised Difference Vegetation Index (NDVI)] with a meteorological parameter [i.e., Standardised Precipitation (std_prec)]. The novel CDI was applied at the Narayani Basin to assess the droughts over the 2004–2013 period, and the results were independently evaluated using streamflow observations to validate the accuracy of the novel drought index. The principal component analysis (PCA) technique was used to determine the contribution of input parameters to the multivariate drought index. The PCA results show a strong positive correlation (0.78) between the CDI and standardised streamflow, indicating the effectiveness of the novel index in monitoring drought conditions. Accordingly, it can be concluded that surface water availability is interdependent on landscape characteristics, such as LST, SC, and NDVI, in addition to the effects of precipitation. Also, the novel CDI can identify the specific drought-affected areas in the Narayani River Basin, offering insights into its drought characteristics beyond traditional drought assessment techniques
SnatchML: hijacking ML models without training access
No full textThe widespread deployment of Machine Learning (ML) models has been accompanied by the emergence of various attacks that threaten their trustworthiness and raise ethical and societal concerns. One such attack is model hijacking, where an adversary seeks to repurpose a victim model to perform a different task than originally intended. Model hijacking can cause significant accountability and security risks since the owner of a hijacked model can be framed for having their model offer illegal or unethical services. Prior works consider model hijacking as a training time attack, whereby an adversary requires full access to the ML model training. In this paper, we consider a stronger threat model for an inference-time hijacking attack, where the adversary has no access to the training phase of the victim model. Our intuition is that ML models, which are typically over-parameterized, might have the capacity to (unintentionally) learn more than the intended task they are trained for. We propose SnatchML, a new training-free model hijacking attack, that leverages the extra capacity learnt by the victim model to infer different tasks that can be semantically related or unrelated to the original one. Our results on models deployed on AWS Sagemaker showed that SnatchML can deliver high accuracy on hijacking tasks. Interestingly, while all previous approaches are limited by the number of classes in the benign task, SnatchML can hijack models for tasks that contain more classes than the original. We explore different methods to mitigate this risk; We propose meta-unlearning, which is designed to help the model unlearn a potentially malicious task while training for the original task. We also provide insights on over-parametrization as a possible inherent factor that facilitates model hijacking, and accordingly, we propose a compression-based countermeasure to counteract this attack. We believe this work offers a previously overlooked perspective on model hijacking attacks, presenting a stronger threat model and higher applicability in real-world contexts. Our code is available at https://github.com/ihsenLab/SnatchML
Analytical Bergsonism
No full textThis paper discusses the phenomenon of “analytical appropriations;” movements based around analytic translations of non-analytic thinkers or traditions. It uses this discussion to propose a new instance of such appropriations: Analytical Bergsonism. Recent years have seen a notable increase in engagements between analytic philosophers and the philosophy of Henri Bergson (1859–1941), who has generally been considered as standing outside the analytic tradition. During and Miquel (2020, 17-42) have recently suggested that we should begin to take seriously the possibility of an “analytic Bergson.” However, no definition or metaphilosophical discussion of what this is supposed to be has been provided. This paper discusses similarities between three analogous analytic appropriations of non-analytic thinkers or systems of thought (Analytic Feminism, Analytical Thomism, and Analytic Theology), and uses an overview of their general structures to provide a descriptive and normative groundwork for Analytical Bergsonism. The paper also addresses several objections against the movement.<br/
Liquid biopsy to identify Barrett’s oesophagus, dysplasia and oesophageal adenocarcinoma: the EMERALD multicentre study
No full textBACKGROUND: There is no clinically relevant serological marker for the early detection of oesophageal adenocarcinoma (EAC) and its precursor lesion, Barrett's oesophagus (BE).OBJECTIVE: To develop and test a blood-based assay for EAC and BE.DESIGN: Oesophageal MicroRNAs of BaRRett, Adenocarcinoma and Dysplasia (EMERALD) was a large, international, multicentre biomarker cohort study involving 792 patient samples from 4 countries (NCT06381583) to develop and validate a circulating miRNA signature for the early detection of EAC and high-risk BE. Tissue-based miRNA sequencing and microarray datasets (n=134) were used to identify candidate miRNAs of diagnostic potential, followed by validation using 42 pairs of matched cancer and normal tissues. The usefulness of the candidate miRNAs was initially assessed using 108 sera (44 EAC, 34 EAC precursors and 30 non-disease controls). We finally trained a machine learning model (XGBoost+AdaBoost) on RT-qPCR results from circulating miRNAs from a training cohort (n=160) and independently tested it in an external cohort (n=295).RESULTS: After a strict process of biomarker discovery and selection, we identified six miRNAs that were overexpressed in all sera of patients compared with non-disease controls from three independent cohorts of different nationalities (miR-106b, miR-146a, miR-15a, miR-18a, miR-21 and miR-93). We established a six-miRNA diagnostic signature using the training cohort (area under the receiver operating characteristic curve (AUROC): 97.6%) and tested it in an independent cohort (AUROC: 91.9%). This assay could also identify patients with BE among patients with gastro-oesophageal reflux disease (AUROC: 94.8%, sensitivity: 92.8%, specificity: 85.1%).CONCLUSION: Using a comprehensive approach integrating unbiased genome-wide biomarker discovery and several independent experimental validations, we have developed and validated a novel blood test that might complement screening options for BE/EAC.TRIAL REGISTRATION NUMBER: NCT06381583</p
Do global COVOL and geopolitical risks affect clean energy prices? Evidence from explainable artificial intelligence models
No full textWe investigate the impact of global common volatility and geopolitical risks on clean energy prices. Our study utilizes daily data from January 1, 2001, to March 18, 2024. Using a new framework based on explainable artificial intelligence (XAI) methods, our findings demonstrate that the COVOL index outperforms the geopolitical risk index in accurately predicting clean energy prices. Furthermore, the Extreme Trees algorithm shows superior performance compared to traditional regression techniques. Our findings indicate that XAI improves transparency, thereby making a substantial contribution to agile decision-making in predicting clean energy prices. Practitioners, including investors and portfolio managers, can enhance investment decisions and manage systemic risks by incorporating COVOL into their risk assessment and asset allocation models.<br/
Non-adherence to immunosuppressive medications in kidney transplant recipients- a systematic scoping review
No full textBackgroundRejection and graft failure remain common in kidney transplant recipients. Non-adherence to immunosuppressive medications is considered a major contributary factor to reduced long-term graft survival, particularly in younger people. Improvements in clinical practice based on adherence studies has been minimal.MethodsJoanna Briggs' Institute Methodology was used. MedlineALL, Embase, Web of Science Core Collection and Scopus databases were searched from January 2000 through to December 2023. Abstract and full text reviews were undertaken independently by two reviewers. Data was collated using a pre-designed extraction tool.Results359 articles met the inclusion criteria. Non-adherence was commonly defined using self-reported questionnaires or pharmacy re-fill rates. Prevalence of non-adherence varied widely. There was little correlation between method of measurement and reported rates of non-adherence. Despite younger age being identified as a risk factor for non-adherence, pooled reported prevalence did not differ significantly in studies reporting prevalence in children, adolescents, or young adults vs. older adults (36.0 % vs. 34.0 %). Interventional studies to detect or improve adherence are highly heterogenous, often report small effects and are limited by the lack of gold-standard methods to measure adherence.DiscussionThis scoping review outlines the complexities of non-adherence to immunosuppressive medications among kidney transplant recipients, highlighting significant variability in adherence definitions, measurements, and intervention efficacy. Reported non-adherence rates vary widely (2–89 %), underscoring the need for standardisation of the definition of non-adherence in research. Findings suggest that non-adherence to immunosuppressive medication is driven by a mix of demographic, psychosocial, and transplant-specific factors. Future research should prioritise standardised definitions of adherence, validated tools to measure adherence, and focus on clinically significant outcomes in non-adherent populations to develop meaningful, impactful interventions for long-term patient benefit.<br/
Instruction and guidance in healthcare simulation: a scoping review
No full textIntroductionThere is growing evidence that instruction and guidance during simulation engagement can enhance explicit and subtle procedural knowledge and skills, medical knowledge, situation awareness and organization, and observation and reflection. However, instruction and guidance to scaffold learners during simulation engagement receive limited attention in published peer-reviewed literature, simulation practice guidelines and instructional design practices. This scoping review aimed to identify specific instruction or guidance strategies used to scaffold learners during simulation engagement, who or what provided support and guidance, who received instruction or guidance, and for what reasons.MethodsGuided by Reiser and Tabak’s perspectives on scaffolding, we conducted a scoping review following JBI Guidance. Included databases were PubMed, CINAHL, Embase, PsycINFO and Web of Science. No date boundary was set. All languages were eligible. Hand searching included six healthcare simulation journals, yielding 9232 articles at the start. Using Covidence, two reviewers independently screened all articles (title and abstract, full-text). Two independent reviewers extracted every third article. The content analysis enabled categorization and frequency counts.ResultsNinety articles were included. A human or computer tutor or a combination of human and computer tutors provides instruction and guidance. Strategies employed by human tutors were verbal guidance, checklists, collaboration scripts, encouragement, modelling, physical guidance and prescribed instructional strategies (e.g., rapid cycle deliberate practice). Strategies employed by computer tutors were audio prompts, visualization, modelling, step-by-step guides, intelligent tutoring systems and pause buttons. Most studies focused on pre-licensure and immediate post-graduate learners but continuing professional development learners were also represented. The most common reason for including instruction and guidance was to enhance learning without specific language regarding how or what aspects of learning were intended to be enhanced.ConclusionAlthough less prominent than pre- and post-simulation instructional strategies (e.g., pre-briefing, debriefing), there is a growing body of literature describing instruction and guidance for scaffolding learners during simulation engagements. Implications for practice, professional guidelines and terminology are discussed
A practical approach to quantitative analytical surface-enhanced Raman spectroscopy
No full textMany of the features of SERS, such as its high sensitivity, molecular specificity and speed of analysis make it attractive as an analytical technique. However, SERS currently remains a specialist technique which has not yet entered the mainstream of analytical chemistry. Therefore, this review draws out the underlying principles for analytical SERS and provides practical tips and tricks for SERS quantitation. The aim is to show the readers how to rationally design their SERS experiments to improve quantitation performance. We begin by introducing the three core components in SERS analysis: (1) the enhancing substrate material, (2) the Raman instrument and (3) the processed data that is used to establish a calibration curve. This is followed by discussion of the analytical figures of merit relevant to SERS. In the following sections each of the three essential components in SERS quantitation and how they affect the quality of the analysis are described in more detail using examples from the literature. Finally, we highlight the current challenges in applying SERS to the analysis of complex real-life samples and briefly introduce the state-of-the-art developments on multifunctional substrates, digital SERS and AI-assisted data processing, which will help SERS rise to the challenge of moving out into routine real-world analysis
DIP-ECOD: improving anomaly detection in multimodal distributions
No full textAnomaly detection algorithms identify unusual events and outliers in large datasets where manual approaches are highly impractical. Most prior anomaly detection methods assume simple unimodal Gaussian data distributions; however, they produce suboptimal results on complex multimodal distributions. To address this problem, we propose DIP-ECOD, a novel anomaly detection algorithm leveraging unsupervised machine learning that generalises to both multimodal and unimodal distributions. DIP-ECOD integrates a dip test within the ECOD framework, using SkinnyDip to split a probability distribution into separate modes, after which ECOD is applied. In this way, difficult-to-find outliers between modes and hidden in the distribution tails of each mode are also detected. Experiments using nine benchmark datasets across a range of domains such as healthcare and imagery demonstrate DIP-ECOD’s improved performance over ECOD in detecting outliers in both multimodal and unimodal distributions, with DIP-ECOD achieving an average AUC score of 0.791 compared to ECOD’s 0.761. Further, using a proprietary enterprise dataset, we show DIP-ECOD effectively identifies anomalous Github commits, indicating its applicability to information security and software vulnerability, where multi modal distributions are expected.<br/