Subset Of Data Research Articles

Predicting saturated hydraulic conductivity from particle size distributions using machine learning

Abstract Estimating saturated hydraulic conductivity $$K_f$$ from particle size distributions (PSD) is very common with empirical formulas, while the use of machine learning for that purpose is not yet widely established. We evaluate the predictive power of six machine learning algorithms, including tree-based, regression-based and network-based methods in estimating $$K_f$$ from the PSD solely. We use a dataset of 4600 samples from the shallow Dutch subsurface for training and testing. The extensive dataset provides not only PSD, but also measured conductivities from permeameter tests. Besides training and testing on the entire data set, we apply the six algorithms to data subsets for the soil types sand, silt and clay. We further test different feature/target-variable combinations such as reducing the input to PSD-derived grain diameters $$d_{10}$$ , $$d_{50}$$ and $$d_{60}$$ or estimating porosity from PSD. We test feature importance and compare results to $$K_f$$ estimates from a selection of empirical formulas. We find that all algorithm can estimate $$K_f$$ from PSD at high accuracy (up to $$R^2/NSE$$ of 0.89 for testing data and 0.98 for the entire data set) and outperform empirical formulas. Particularly, tree-based algorithms are well suited and robust. Reducing information in the feature variables to grain diameters works well for predicting $$K_f$$ of sandy samples, but is less robust for silt and clay rich samples. $$d_{10}$$ also shows to be the most influential feature here. An interesting, but not surprising outcome is that PSD is not a suitable predictor for porosity. Overall, our results confirm that machine learning algorithms are a powerful tool for determining $$K_f$$ from PSD. This is promising for applications to e.g. deep-drilling data sets or low-effort and robust $$K_f$$ -estimation of single samples.

Merging textual and numerical databases: a steppingstone for statistical analyses of illegal events

Abstract This paper aims to define a methodological path—merging judgments and official statistical data—to organize complete, objective, and reliable data in a database, thus simplifying the analysis of illegal social phenomena. Judiciary judgments are a new data source: they deal with illegal events that describe social phenomena—even if they are only the "illegal" ones—and contain objective and reliable data and information. Judiciary judgments are also texts, so the first step is a statistical textual analysis and text mining techniques to discover information and organize it in a statistical database. The final database is obtained by integrating numerical data from other information sources. It therefore has statistical properties such as reliability, completeness and updating. Subsequent statistical analyses or modelling are then possible based on the entire set or subsets of data adequately extracted from the implemented statistical database. We present some results obtained from judgments about corruption in order to demonstrate the advantages of linking textual databases (textual analyses on judgments) and numerical databases (from ISTAT). The proposed methodology can benefit different stakeholders, such as researchers, policymakers, and other enforcement actors. It is independent of the specific software used and remains valid when applied to other illegal activities (e.g., organized crime, tax crime, and money laundering). Furthemore, the results may be even more effective if the institutional actors involved have access to judgments at all levels, thus overcoming potential privacy concerns. The methodology could also be used to support evidence-based policy in the fight against crime and illegal activities.

The Association Between COVID-19 Vaccination Uptake and Information-Seeking Behaviors Using the Internet: Nationwide Cross-Sectional Study.

The COVID-19 pandemic, declared in March 2020, profoundly affected global health, societal, and economic frameworks. Vaccination became a crucial tactic in combating the virus. Simultaneously, the pandemic likely underscored the internet's role as a vital resource for seeking health information. The proliferation of misinformation on social media was observed, potentially influencing vaccination decisions and timing. This study aimed to explore the relationship between COVID-19 vaccination rates, including the timing of vaccination, and reliance on internet-based information sources in Japan. Using a cross-sectional study design using a subset of panel data, this nationwide survey was conducted in 7 waves. A total of 10,000 participants were randomly selected through an internet survey firm, narrowing down to 8724 after applying inclusion and exclusion criteria. The primary outcome was the COVID-19 vaccination date, divided into vaccinated versus unvaccinated and early versus late vaccination groups. The main exposure variable was the use of internet-based information sources. Control variables included gender, family structure, education level, employment status, household income, eligibility for priority COVID-19 vaccination due to pre-existing medical conditions, and a health literacy scale score. Two regression analyses using generalized estimating equations accounted for prefecture-specific correlations, focusing on vaccination status and timing. In addition, chi-square tests assessed the relationship between each information source and vaccination rates. Representing a cross-section of the Japanese population, the regression analysis found a significant association between internet information seeking and higher vaccination rates (adjusted odds ratio [aOR] 1.42 for those younger than 65 years; aOR 1.66 for those aged 65 years and older). However, no significant link was found regarding vaccination timing. Chi-square tests showed positive associations with vaccination for television, government web pages, and web news, whereas blogs and some social networking sites were negatively correlated. Internet-based information seeking is positively linked to COVID-19 vaccination rates in Japan, underscoring the significant influence of online information on public health decisions. Nonetheless, certain online information sources, including blogs and some social networks, negatively affected vaccination rates, warranting caution in their use and recognition. The study highlights the critical role of credible online sources in public health communication and the challenge of combating misinformation on less regulated platforms. This research sheds light on how the digital information landscape influences health behaviors, stressing the importance of accurate and trustworthy health information amidst global health emergencies.

Digital forgetting in large language models: a survey of unlearning methods

Large language models (LLMs) have become the state of the art in natural language processing. The massive adoption of generative LLMs and the capabilities they have shown have prompted public concerns regarding their impact on the labor market, privacy, the use of copyrighted work, and how these models align with human ethics and the rule of law. As a response, new regulations are being pushed, which require developers and service providers to evaluate, monitor, and forestall or at least mitigate the risks posed by their models. One mitigation strategy is digital forgetting: given a model with undesirable knowledge or behavior, the goal is to obtain a new model where the detected issues are no longer present. Digital forgetting is usually enforced via machine unlearning techniques, which modify trained machine learning models for them to behave as models trained on a subset of the original training data. In this work, we describe the motivations and desirable properties of digital forgetting when applied to LLMs, and we survey recent works on machine unlearning. Specifically, we propose a taxonomy of unlearning methods based on the reach and depth of the modifications done on the models, we discuss and compare the effectiveness of machine unlearning methods for LLMs proposed so far, and we survey their evaluation. Finally, we describe open problems of machine unlearning applied to LLMs and we put forward recommendations for developers and practitioners.

A test of target congruence theory on cyber dating abuse among youth

While a number of studies have provided support for target congruence theory in explaining various forms of physical victimization, researchers have started to examine the efficacy of the theory on cyber victimization, specifically cyberstalking. However, no study to date has examined the applicability of this theory to cyber abuse victimization among youth in dating relationships. To this end, the current study applied target congruence theory to examine the influence of target vulnerability, target gratifiability, and target antagonism on cyber dating abuse victimization experienced by high school youth. A subset of data obtained from the Technology, Teen Dating Violence and Abuse, and Bullying in Three States (n = 2,250) showed support for the theory. More specifically, two out of the three concepts used to capture target vulnerability (depression and drug use) and two out of the three measures of target antagonism (hostility and manipulation/fear) were significant in predicting cyber dating abuse victimization, while target gratifiability (sexual cyber perpetration) was also significant. An overall review of the results, as well as the study’s limitations and suggestions for future research, are discussed.

Integration of PRISMA Hyperspectral Satellite Data with Ground Based Geological investigation for Mapping alteration minerals associated with the Neem-Ka-Thana Cu belt in Rajasthan, India

Hydrothermal deposits are commonly associated with specific alteration minerals that serve as key indicators for mineral exploration. The Neem Ka Thana Cu Belt, situated southeast of the Khetri Cu deposit within the Alwar-Ajabgarh sub-basin of the North Delhi Fold Belt, is notable for its Bornite-rich Cu-S mineralization. Despite its geological significance, detailed spectral mapping to delineate the alteration minerals associated with base metal mineralisation remained limited. This study addresses this gap by utilizing the “PRecursore IperSpettrale della Missione Applicativa” (PRISMA) hyperspectral sensor to detect and map alteration minerals associated with Cu-S mineralization.To achieve this, we applied Relative Band Depth (RBD) indices on targeted spectral subsets of PRISMA data to identify Fe-oxides/hydroxides and Al-OH-bearing minerals. We detected key alteration minerals, including muscovite, illite, chlorite, montmorillonite and Fe-oxide and hydroxides such as goethite, hematite, and limonite, by targeting their diagnostic absorption features. The resulting spectral maps highlighting the spatial distribution of the targeted mineral groups were validated with field investigations and laboratory assessments. The study demonstrates that the integration of hyperspectral analysis with conventional geological techniques can help to understand the mineral distribution and associated alteration processes. The use of PRISMA hyperspectral data provides a powerful, non-invasive means for reconnaissance mapping of exposed lithologies, delivering targeted information that is crucial for optimizing subsequent field investigations and drilling operations. The present work highlights the potential of PRISMA data in advancing the methodologies of mineral exploration and lithological mapping, contributing valuable insights for the geoscientific community.

Repeatable process for extracting health data from HL7 CDA documents.

This study aims to address the gap in the literature on converting real-world Clinical Document Architecture (CDA) data into the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM), focusing on the initial steps preceding the mapping phase. We highlight the importance of a repeatable Extract-Transform-Load (ETL) pipeline for health data extraction from HL7 CDA documents in Estonia for research purposes. We developed a repeatable ETL pipeline to facilitate the extraction, cleaning, and restructuring of health data from CDA documents to OMOP CDM, ensuring a high-quality and structured data format. This pipeline was designed to adapt to continuously updated data exchange format changes and handle various CDA document subsets for different scientific studies. We demonstrated via selected use cases that our pipeline successfully transformed a significant portion of diagnosis codes, body weight and eGFR measurements, and PAP test results from CDA documents into OMOP CDM, showing the ease of extracting structured data. However, challenges such as harmonising diverse coding systems and extracting lab results from free-text sections were encountered. The iterative development of the pipeline facilitated swift error detection and correction, enhancing the process's efficiency. After a decade of focused work, our research has led to the development of an ETL pipeline that effectively transforms HL7 CDA documents into OMOP CDM in Estonia, addressing key data extraction and transformation challenges. The pipeline's repeatability and adaptability to various data subsets make it a valuable resource for researchers dealing with health data. While tested on Estonian data, the principles outlined are broadly applicable, potentially aiding in handling health data standards that vary by country. Despite newer health data standards emerging, the relevance of CDA for retrospective health studies ensures the continuing importance of this work.

What’s inside the box? Approximating containerised trade flows based on trade commodity values

ABSTRACT Containerisation plays a critical role in global supply chains and their container flows. The United Nations COMTRADE database provides data on trade flows across various product categories and time periods that are likely to be containerised. This paper undertakes a detailed analysis of the corresponding data, aiming to evaluate whether the data assumptions generally used provide a good approximation of actual containerised trade flows. An iterative and randomised assessment of different COMTRADE data subsets is undertaken. The analysis shows the assumption that the same product categories are likely to be containerised at a similar level for all countries is invalid. This paper seeks to provide accurate and tested insights on the nature of containerised freight.

IMPACT OF PREPROCESSING AND COMPARISON OF NEURAL NETWORK ENSEMBLE METHODS FOR SEGMENTATION OF THE THORACIC SPINE IN X-RAY IMAGES

Context. Automatic segmentation of medical images plays an important role in the process of automating the detection of various diseases in the spine and the use of radiography is the most accessible means of predicting diseases. Over the years many studies have been conducted on the topic of image segmentation. One of the many methods for improving image segmentation is the use of neural network ensembles. Objective. The aims of this study were to investigate the impact of preprocessing and compare the main methods of neural network ensembles and their effect on the segmentation of the thoracic region, in this study the area was considered which consists of the vertebrae: Th8, Th9, Th10, Th11. Method. To begin with, the influence of preprocessing of X-ray images was considered, which included the following methods: histogram equalization for contrast enhancement, contrast-limited adaptive histogram equalization, logarithmic transform method, median filter, Gaussian filter, and bilateral filter. To study the influence of neural network ensemble on segmentation quality, several methods were used. Averaging method – a simple half-averaging method. Weighted averaging method – an improved version of the averaging method which uses weights for each network, the higher the network weight, the greater its influence on averaging. Method of cumulative averaging – a modified averaging method in which each ensemble receives an averaged image, after which all the results of the ensembles are averaged. Bagging – method of averaging networks trained on different data, n networks are used, the training sample is divided into n parts, and each neural network is trained on its own subset of data, as a result, the averaging method is used for predictions. Averaging method for a large number of networks – in this method, 100 neural networks were trained, after which the averaging method was used. Method of averaging mask shapes – this method uses a distance transform to average multiple masks into one shape average. Results. It was investigated that the use of different methods of image preprocessing does not guarantee an improvement in the quality of segmentation of the spine region on X-ray images, but even on the contrary worsens the quality of segmentation. Different methods of combining predictions of neural network ensembles were considered, which made it possible to find out the pros and cons of specific methods for the task of segmentation of X-ray images. Conclusions. The experiments conducted allowed us to conclude that the use of any preprocessing methods should not be used for segmentation of X-ray images. Also, due to a large number of architectures and methods for combining predictions, the behavior of ensemble methods was studied, which will allow us to further determine the necessary approach for segmentation of X-ray images. Further study of the weighted averaging method and the mask shape averaging method will make it possible to improve the obtained result and achieve even greater success in segmentation.

An Approach for Detecting Parkinson’s Disease by Integrating Optimal Feature Selection Strategies with Dense Multiscale Sample Entropy

Parkinson’s disease (PD) is a neurological disorder that severely affects motor function, especially gait, requiring accurate diagnosis and assessment instruments. This study presents Dense Multiscale Sample Entropy (DM-SamEn) as an innovative method for diminishing feature dimensions while maintaining the uniqueness of signal features. DM-SamEn employs a weighting mechanism that considers the dynamic properties of the signal, thereby reducing redundancy and improving the distinctiveness of features extracted from vertical ground reaction force (VGRF) signals in patients with Parkinson’s disease. Subsequent to the extraction process, correlation-based feature selection (CFS) and sequential backward selection (SBS) refine feature sets, improving algorithmic accuracy. To validate the feature extraction and selection stage, three classifiers—Adaptive Weighted K-Nearest Neighbors (AW-KNN), Radial Basis Function Support Vector Machine (RBF-SVM), and Multilayer Perceptron (MLP)—were employed to evaluate classification efficacy and ascertain optimal performance across selection strategies, including CFS, SBS, and the hybrid SBS-CFS approach. K-fold cross-validation was employed to provide improved evaluation of model performance by assessing the model on various data subsets, thereby mitigating the risk of overfitting and augmenting the robustness of the results. As a result, the model demonstrated a significant ability to differentiate between PD patients and healthy controls, with classification accuracy reported as ACC [CI 95%: 97.82–98.5%] for disease identification and ACC [CI 95%: 96.3–97.3%] for severity assessment. Optimal performance was primarily achieved through feature sets chosen using SBS and the integrated SBS-CFS methods. The findings highlight the model’s potential as an effective instrument for diagnosing PD and assessing its severity, contributing to advancements in clinical management of the condition.

Review of methodology and re-analysis of lipidomic data focusing on specialised pro-resolution lipid mediators (SPMs) in a human model of resolving inflammation.

Using a model of UV-killed E. coli driven dermal inflammation in healthy human volunteers, we originally reported that following inflammatory resolution there was infiltration of macrophages, which, through prostanoids including prostaglandin (PG) E2, imprints long-term tissue immunity. In addition to the prostanoids, data on levels of Specialised Pro-Resolution Lipid Mediators (SPMs) throughout inflammatory onset, resolution and post-resolution phases of this model were presented, but as illustrations rather than as primary data. Therefore, in response to a request for increased transparency, a subset of the original data from our human UV-killed E. coli model was re-analysed by two experts from an independent laboratory alongside a review of the methodology used. The prostanoids were detected robustly following re-analysis but the areas of the chromatographic peaks of the SPM lipid mediators were too small to yield amounts that could be reliably detected and/or quantified using community standards. Importantly, with prostanoids including PGE2 being robustly detected, this re-analysis does not alter the original report that post-resolution PGs imprint tissue immunity. Here we show the outcome of this re-analysis and review the biology surrounding SPMs as a result.

E-cigarette (EC) and heated tobacco product (HTP) use in the United Arab Emirates, an emerging EC and HTP market: A cross-sectional analysis of the International Tobacco Control (ITC) UAE Survey.

E-cigarettes (ECs) and heated tobacco products (HTPs) are recent arrivals to the nicotine product market in the Middle East, which are rapidly growing in popularity in the region. There is a lack of surveillance data at the country-level on use of these products and factors associated with their use. This study analyzed a subset of data from the UAE Healthy Future Study, a population-based cohort study of the Emirati population, to determine the factors associated with EC and HTP use among a sample of Emirati adults (≥ 18 years). The baseline assessment and supplementary questionnaires, conducted from 2016 to 2023, included data on combustible tobacco use, EC, and HTP use and sociodemographic characteristics. Of the 2,041 individuals who answered questions on EC use, 32% reported ever using them. Of 521 people providing data on HTP, 30% reported ever using them. After adjusting for age, sex, education, perceived harms and perceived addictiveness of EC, current EC use was associated with baseline combustible tobacco smoking (aOR = 27.63, 95% confidence interval [CI] 14.39, 53.06), users of a younger age (aOR=0.91, 95%CI 0.88, 0.95), and users of male sex (OR = 2.15, 95% CI 1.21, 3.81). Current HTP use was less common, but was also associated with baseline combustible tobacco use. Use of ECs and HTPs was more common among those who used combustible tobacco. Future research should examine use trajectories among those who do and do not smoke, as well as uptake of these products among youth. Non-combustible nicotine products are growing in popularity in the Middle East Region. Our study found that EC and HTP use is associated with baseline combustible tobacco use and that concurrent users may use them to cut down on their combustible tobacco use. Continued comprehensive population-based monitoring of all tobacco and nicotine products, especially EC and HTP use, will provide current data to aid in appropriately informing public health and harm reduction messages and programming.

Online variational Gaussian process for time series data

Gaussian processes (GPs) are a powerful and popular framework for addressing machine learning problems, particularly for time-dependent data such as that generated by the Internet of Things (IoT). GPs offer a compelling choice for constructing real-valued nonlinear models due to their inherent flexibility and ability to quantify uncertainty. However, traditional GP methods are often hindered by cubic computational complexity, making them impractical for the massive and potentially unbounded datasets commonly encountered in IoT applications. To address this issue, researchers have developed various sparse approximation methods that significantly reduce the computational burden of GPs. Among these, pseudo-point approximations have proven to be highly influential, leveraging a subset of the training data to represent the entire observation space. The variational sparse GP is a state-of-the-art approach that approximates the posterior distribution of GP models, enabling faster and more efficient predictions in real-time data series. However, integrating variational inference into the GP framework for sequentially arriving data remains a significant challenge, particularly when dealing with time series data where the underlying data distribution evolves over time. In this paper, we propose the OnLine Variational Gaussian Process (OLVGP) algorithm, which introduces a novel approach for dynamically managing the number of inducing points based on the concept of eigenfunction inducing features. Unlike traditional methods that rely on a fixed number of inducing points, OLVGP adaptively adjusts the number of inducing points as new data arrives and optimizes them from the model, ensuring that the model remains computationally efficient while maintaining high predictive accuracy. Our method capitalizes on the sophisticated design of the online variational inference framework, ensuring a technically robust derivation and implementation. We validate the effectiveness and efficiency of our approach through both synthetic examples and real-world experiments. The results demonstrate that OLVGP not only substantially reduces computational costs compared to traditional sparse GP methods but also dynamically adapts to the evolving data, delivering improved performance in time series prediction.

Distributed System Identification for Linear Stochastic Systems Under an Adaptive Event‐Triggered Scheme

ABSTRACTThis article considers a distributed identification problem for linear stochastic systems whose input and output observations are scheduled by an adaptive event‐triggered scheme. An event detector with time‐varying thresholds is designed to control the transmission of measurements from the sensors to the estimators, which leads to that only a subset of input and output data is available for identification. The estimators exchange information over a network and cooperatively identify the unknown parameters. A distributed recursive identification algorithm under the event‐triggered scheme is proposed based on the distributed stochastic approximation algorithm with expanding truncations (DSAAWET). Under mild assumptions, the strong consistency of the algorithm is proved, that is, the estimates generated from each estimator achieve consensus and converge to the true parameters with probability one. Finally, two numerical examples are provided to validate the theoretical results of the algorithm.

Effect of patient-contextual skin images in human- and artificial intelligence-based diagnosis of melanoma: Results from the 2020 SIIM-ISIC melanoma classification challenge.

While the high accuracy of reported AI tools for melanoma detection is promising, the lack of holistic consideration of the patient is often criticized. Along with medical history, a dermatologist would also consider intra-patient nevi patterns, such that nevi that are different from others on a given patient are treated with suspicion. To evaluate whether patient-contextual lesion-images improves diagnostic accuracy for melanoma in a dermoscopic image-based AI competition and a human reader study. An international online AI competition was held in 2020. The task was to classify dermoscopy images as melanoma or benign lesions. A multi-source dataset of dermoscopy images grouped by patient were provided, and additional use of public datasets was permitted. Competitors were judged on area under the receiver operating characteristic (AUROC) on a private leaderboard. Concurrently, a human reader study was hosted using a subset of the test data. Participants gave their initial diagnosis of an index case (melanoma vs. benign) and were then presented with seven additional lesion-images of that patient before giving a second prediction of the index case. Outcome measures were sensitivity and specificity. The top 50 of 3308 AI competition entries achieved AUROC scores ranging from 0.943 to 0.949. Few algorithms considered intra-patient lesion patterns and instead most evaluated images independently. The median sensitivity and specificity of human readers before receiving contextual images were 60.0% and 86.7%, and after were 60.0% and 85.7%. Human and AI algorithm performance varied by image source. This study provided an open-source state-of-the-art algorithm for melanoma detection that has been evaluated at multiple centres. Patient-contextual images did not positively impact performance of AI algorithms or human readers. Providing seven contextual images and no total body image may have been insufficient to test the applicability of the intra-patient lesion patterns.

Disabled Students in Health and Social Services Fieldwork: Perceptions of Canadian Fieldwork Educators and Academic Coordinators

Ensuring equitable access to professional education programs for learners who need accommodations is distinctly challenging when education moves beyond the classroom into clinical or fieldwork sites. Fieldwork educators and university academic coordinators who arrange fieldwork placements work with university accessibility services and students to arrange required accommodations, while preserving confidentiality, maintaining high learning standards, and ensuring attainment of professional competencies. This work is complicated by time pressures and heavy caseloads in fieldwork settings. Here we report on a subset of data from a cross-Canada online survey of fieldwork educators (n = 233) and academic coordinators (n = 54) in 10 health and social service professions. Using descriptive statistics, we analyze responses to two question series concerning perceptions of the capacity of disabled students to attain professional competencies, and overall perceptions of students who need accommodations. Respondents showed most concern about competency attainment for learners with cognitive or learning disabilities, followed by neurological and mental health issues. Thematic analysis of open-ended comments suggests doubt regarding the ability of institutional fieldwork sites to adequately implement accommodations. In their perception of learners who need accommodations, academic coordinators were somewhat more negative than fieldwork educators, in particular seeing students who need accommodations as a potential burden that could harm placement relationships with fieldwork sites. They tended to indicate that fieldwork success depended on student insight and self-advocacy. Struggles faced by disabled students in health and social service professions appear to be occasioned not only by disabling systems and institutions, but also by perceptions that they may have diminished competence.

The Need for a Whole of Family Response: The Invisibility of Siblings in Adolescent-to-Parent Violence Interventions

Abstract Purpose Adolescent-to-parent violence (APV), especially son-to-mother violence, often occurs in the context of adult-perpetrated domestic and family violence (DFV). APV interventions largely focus on the behaviour of the young person using violence, and their relationship with the protective parent, often the mother. Focus on the mother-young person dyad is important yet ignores their relationships with siblings, who have also experienced DFV and APV. If dynamics within the home are unaccounted for, strategies developed within the contained environment of an intervention program are unlikely to be sustained. This paper examines the experiences of siblings whose mothers and brothers are attending an APV intervention. In order to inform more responsive policies and practices, it addresses two research questions: 1) What are the needs of siblings whose brothers are engaging in APV? and (2) how can APV interventions better recognise and respond to the needs of siblings? Methods This paper reports on a subset of qualitative data from the evaluation of an Australian APV intervention, conducted between 2020 and 2022. Interviews were undertaken with mothers (n = 17), sons using APV (n = 9), practitioners (n = 11) and a sibling (n = 1). Case files (n = 16) containing data related to the levels of violence witnessed by siblings were also reviewed. Results Thematic analysis produced five themes: (1) intersecting complexity; (2) recognition of violence experienced by siblings; (3) different sibling cohorts; (4) family disruption; and (5) siblings and the program. Conclusion We argue that a whole of family response is needed for families recovering from DFV where APV is present.

Hierarchical Bayesian modeling for Inverse Uncertainty Quantification of system thermal-hydraulics code using critical flow experimental data

The best estimate plus uncertainty methodology in nuclear system thermal-hydraulic studies necessitates a comprehensive understanding of uncertainties in system code predictions. The forward uncertainty quantification (UQ) process involves the propagation of input uncertainties through the computational models to obtain uncertainties in the outputs. To this end, achieving an accurate estimation of input uncertainties is important, which is the focus of inverse UQ (IUQ). Traditionally, research in Bayesian IUQ within the nuclear engineering domain has largely relied on single-level Bayesian inference. While being effective for relatively small datasets, this approach encounters limitations for cases with large datasets. The use of a single-level model may prove inefficient, as the resultant posterior distributions can significantly differ when distinct subsets of data are employed. To address this issue, we employ an hierarchical Bayesian model for IUQ. This approach involves organizing observations into different groups based on the test conditions, thereby accommodating varying calibration parameters across these distinct groups. In this study, we developed and implemented a hierarchical Bayesian IUQ method to consider the grouping effect of critical flow measurement data from various geometries. Comparing the outcomes of IUQ under different selections of test data using hierarchical Bayesian IUQ against those obtained from single-level Bayesian IUQ, the forward propagation of hierarchical Bayesian IUQ results demonstrates a notably improved agreement with the experimental data.

Classifying active and inactive states of growing rabbits from accelerometer data using machine learning algorithms

Using wearable accelerometers is gaining traction in research and animal production management for monitoring animal behaviour. In this study, the objective was to automatically detect rabbit activity/inactivity states from accelerometer data in growing rabbits. For that purpose, 16 animals were equipped with an accelerometer and filmed for 2 weeks. A total of 10 hours of video across all the rabbits were annotated manually using the Boris software, identifying 6 classes of different behaviours: lying, eating, moving, grooming, walking and drinking which were grouped into two classes: active and inactive. Accelerometer signal and video annotations were manually synchronized. The static and dynamic components of the signal were isolated by applying a low-pass and high-pass filter and 4 additional time series were derived from these components. The signal was segmented into time windows of different sizes: 1, 3, 5, 7 and 9 seconds. For each window, a total of 41 features were extracted in the time and frequency domain. Different subsets of data containing an increasing number (from 5 to 25 in steps of 5) of the most informative features identified with random forest (RF) were used to train a binary classification model (inactive as a positive class). The classification performance of RF, support vector machine (SVM) and gradient boosting (GB) was evaluated. A nested cross-validation (CV) with an outer Leave-One-Animal Out CV and an inner threefold CV for hyperparameter tunning was implemented. The same resampling was implemented for each window size and each classifier so that the models were evaluated with the same data sets. The performance was evaluated on the test datasets using different metrics: precision, recall, F1 score and accuracy. Results showed that the classifiers perform very similarly. With the best configuration (window size of 9 s and with the 5 most important features) the RF model reaches a median precision of 1 (Q1=0.99, Q3=1) and a median recall of 0.93 (Q1=0.89, Q3=0.97). These results showed that the model is highly reliable in correctly classifying positive instances. Additionally, achieving a recall of 0.93 emphasizes the model's effectiveness in capturing a substantial portion of positive instances. Accelerometers combined with machine learning models therefore hold great promise for monitoring rabbit activity and for a range of applications in animal science and behaviours.

Ricci Curvature-Based Graph Sparsification for Continual Graph Representation Learning.

Memory replay, which stores a subset of historical data from previous tasks to replay while learning new tasks, exhibits state-of-the-art performance for various continual learning applications on the Euclidean data. While topological information plays a critical role in characterizing graph data, existing memory replay-based graph learning techniques only store individual nodes for replay and do not consider their associated edge information. To this end, based on the message-passing mechanism in graph neural networks (GNNs), we present the Ricci curvature-based graph sparsification technique to perform continual graph representation learning. Specifically, we first develop the subgraph episodic memory (SEM) to store the topological information in the form of computation subgraphs. Next, we sparsify the subgraphs such that they only contain the most informative structures (nodes and edges). The informativeness is evaluated with the Ricci curvature, a theoretically justified metric to estimate the contribution of neighbors to represent a target node. In this way, we can reduce the memory consumption of a computation subgraph from O(dL) to O(1) and enable GNNs to fully utilize the most informative topological information for memory replay. Besides, to ensure the applicability on large graphs, we also provide the theoretically justified surrogate for the Ricci curvature in the sparsification process, which can greatly facilitate the computation. Finally, our empirical studies show that SEM outperforms state-of-the-art approaches significantly on four different public datasets. Unlike existing methods, which mainly focus on task incremental learning (task-IL) setting, SEM also succeeds in the challenging class incremental learning (class-IL) setting in which the model is required to distinguish all learned classes without task indicators and even achieves comparable performance to joint training, which is the performance upper bound for continual learning.