Discriminant Model Research Articles

SUnSeT: spectral unmixing of hyperspectral images for phenotyping soybean seed traits

Key messageHyperspectral features enable accurate classification of soybean seeds using linear discriminant analysis and GWAS for novel seed trait genes.Evaluating crop seed traits such as size, shape, and color is crucial for assessing seed quality and improving agricultural productivity. The introduction of the SUnSet toolbox, which employs hyperspectral sensor-derived image analysis, addresses this necessity. In a validation test involving 420 seed accessions from the Korean Soybean Core Collections, the pixel purity index algorithm identified seed- specific hyperspectral endmembers to facilitate segmentation. Various metrics extracted from ventral and lateral side images facilitated the categorization of seeds into three size groups and four shape groups. Additionally, quantitative RGB triplets representing seven seed coat colors, averaged reflectance spectra, and pigment indices were acquired. Machine learning models, trained on a dataset comprising 420 accession seeds and 199 predictors encompassing seed size, shape, and reflectance spectra, achieved accuracy rates of 95.8% for linear discriminant analysis model. Furthermore, a genome-wide association study utilizing hyperspectral features uncovered associations between seed traits and genes governing seed pigmentation and shapes. This comprehensive approach underscores the effectiveness of SUnSet in advancing precision agriculture through meticulous seed trait analysis.

AI-based mining of biomedical literature: Applications for drug repurposing for the treatment of dementia.

This manuscript outlines our project involving the application of AGATHA, an AI-based literature mining tool, to discover drugs with the potential for repurposing in the context of neurocognitive disorders. The primary objective is to identify connections between approved medications and specific health conditions through advanced statistical analysis, including techniques like Partial Least Squares Discriminant Analysis (PLSDA) and unsupervised clustering. The methodology involves grouping scientific terms related to different health conditions and genes, followed by building discrimination models to extract lists of disease-specific genes. These genes are then analyzed through pathway analysis to select candidates for drug repurposing.

Online inspection of blackheart in potatoes using visible-near infrared spectroscopy and interpretable spectrogram-based modified ResNet modeling.

Blackheart is one of the most common physiological diseases in potatoes during storage. In the initial stage, black spots only occur in tissues near the potato core and cannot be detected from an outward appearance. If not identified and removed in time, the disease will seriously undermine the quality and sale of theentire batch of potatoes. There is an urgent need to develop a method for early detection of blackheart in potatoes. This paper used visible-near infrared (Vis/NIR) spectroscopy to conduct online discriminant analysis on potatoes with varying degrees of blackheart and healthy potatoes to achieve real-time detection. An efficient and lightweight detection model was developed for detecting different degrees of blackheart in potatoes by introducing the depthwise convolution, pointwise convolution, and efficient channel attention modules into the ResNet model. Two discriminative models, the support vector machine (SVM) and the ResNet model were compared with the modified ResNet model. The prediction accuracy for blackheart and healthy potatoes test sets reached 0.971 using the original spectrum combined with a modified ResNet model. Moreover, the modified ResNet model significantly reduced the number of parameters to 1434052, achieving a substantial 62.71% reduction in model complexity. Meanwhile, its performance was evidenced by a 4.18% improvement in accuracy. The Grad-CAM++ visualizations provided a qualitative assessment of the model's focus across different severity grades of blackheart condition, highlighting the importance of different wavelengths in the analysis. In these visualizations, the most significant features were predominantly found in the 650-750 nm range, with a notable peak near 700 nm. This peak was speculated to be associated with the vibrational activities of the C-H bond, specifically the fourth overtone of the C-H functional group, within the molecular structure of the potato components. This research demonstrated that the modified ResNet model combined with Vis/NIR could assist in the detection of different degrees of black in potatoes.

Altered Metabolism during the Dark Period in Drosophila Short Sleep Mutants.

Sleep is regulated via circadian mechanisms, but effects of sleep disruption on physiological rhythms, in particular metabolic cycling, remain unclear. To examine this question, we probed diurnal metabolic alterations of two Drosophila short sleep mutants, fumin and sleepless. Samples were collected with high temporal sampling (every 2 h) over 24 h under a 12:12 light:dark cycle, and profiling was done using an ion-switching LCMS/MS method. Fewer metabolites with 24 h oscillations were noted with short sleep (50 and 46 in fumin and sleepless, BH. Q < 0.2 by RAIN analysis) compared to a wild-type control (iso31, 63 with BH. Q < 0.2), and peak phases of the sleep mutants were consolidated into two major phase peaks at mid-day and middle of night. Overall, altered nicotinate/nicotinamide, alanine/aspartate/glutamate, acetylcholine, glyoxylate/dicarboxylate, and TCA cycle metabolism were observed in the short sleep mutants, indicative of increased energetic demand and oxidative stress compared to wild type. Both changes in cycling and discriminant models suggest unique alterations in the dark period indicative of constrained metabolic networks. Thus, we conclude that sleep loss alters metabolic function uniquely throughout the day, and further examination of specific mechanisms is warranted.

Clustering single-cell RNA sequencing data via iterative smoothing and self-supervised discriminative embedding.

Single-cell transcriptome sequencing (scRNA-seq) is a high-throughput technique used to study gene expression at the single-cell level. Clustering analysis is a commonly used method in scRNA-seq data analysis, helping researchers identify cell types and uncover interactions between cells. However, the choice of a robust similarity metric in the clustering procedure is still an open challenge due to the complex underlying structures of the data and the inherent noise in data acquisition. Here, we propose a deep clustering method for scRNA-seq data called scRISE (scRNA-seq Iterative Smoothing and self-supervised discriminative Embedding model) to resolve this challenge. The model consists of two main modules: an iterative smoothing module based on graph autoencoders designed to denoise the data and refine the pairwise similarity in turn to gradually incorporate cell structural features and enrich the data information; and a self-supervised discriminative embedding module with adaptive similarity threshold for partitioning samples into correct clusters. Our approach has shown improved quality of data representation and clustering on seventeen scRNA-seq datasets against a number of state-of-the-art deep learning clustering methods. Furthermore, utilizing the scRISE method in biological analysis against the HNSCC dataset has unveiled 62 informative genes, highlighting their potential roles as therapeutic targets and biomarkers.

Mine water inrush source discrimination model based on KPCA-ISSA-KELM.

In order to ensure the safety of coal mine production, a mine water source identification model is proposed to improve the accuracy of mine water inrush source identification and effectively prevent water inrush accidents based on kernel principal component analysis (KPCA) and improved sparrow search algorithm (ISSA) optimized kernel extreme learning machine (KELM). Taking Zhaogezhuang mine as the research object, firstly, Na+, Ca2+, Mg2+, Cl-, SO2- 4 and HCO- 3 were selected as evaluation indexes, and their correlation was analyzed by SPSS27 software, with reducing the dimension of the original data by KPCA. Secondly, the Sine Chaotic Mapping, dynamic adaptive weights, and Cauchy Variation and Reverse Learning were introduced to improve the Sparrow Search Algorithm (SSA) to strengthen global search ability and stability. Meanwhile, the ISSA was used to optimize the kernel parameters and regularization coefficients in the KELM to establish a mine water inrush source discrimination model based on the KPCA-ISSA-KELM. Then, the mine water source data are input into the model for discrimination in compared with discrimination results of KPCA-SSA-KELM, KPCA-KELM, ISSA-KELM, SSA-KELM and KELM models. The results of the study show as follows: The discrimination results of the KPCA-ISSA-KELM model are in agreement with the actual results. Compared with the other models, the accuracy of the KPCA-ISSA-KELM model is improved by 8.33%, 12.5%, 4.17%, 21.83%, and 25%, respectively. Finally, when these models were applied to discriminate water sources in a coal mine in Shanxi, and the misjudgment rates of each model were 28.57%, 19.05%, 14.29%, 23.81%, 9.52% and 4.76%, respectively. From this, the KPCA-ISSA-KLEM model is the most accurate about discrimination and significantly better than other models in other evaluation indicators, verifying the universality and stability of the model. It can be effectively applied to the discrimination of inrush water sources in mines, providing important guarantees for mine safety production.

Gut microbial signatures of patients with diarrhea-predominant irritable bowel syndrome and their healthy relatives

Abstract Aims Irritable bowel syndrome (IBS) is a prevalent gastrointestinal disorder, encompassing diarrhea-predominant irritable bowel syndrome (IBS-D). Here, we utilized 16S rDNA gene sequencing to identify potential microbial drivers of IBS-D. Methods and Results A total of 30 healthy relatives and 27 patients with IBS-D were recruited. Clinical data and fecal samples were collected from patients and controls. 16S rDNA gene sequencing was performed to obtain fecal bacterial data. Differences in community composition were evaluated utilizing analysis of similarity (ANOSIM) using Bray–Curtis dissimilarity. The Wilcoxon rank sum test was used to compare differences in taxa and functional pathways. Finally, the key gut microbiota was identified using the random forest algorithm. Gut microbiota diversity, estimated through the Observe, Chao1, and abundance-based coverage estimator (ACE) indices, was significantly lower in the IBS-D patients than in the healthy relatives. ANOSIM analysis further confirmed significant differences in the composition of the gut microbiota between IBS-D patients and healthy relatives, with an R value of 0.106 and a P-value of 0.005. Notably, the IBS-D patients exhibited a significant enrichment of specific bacterial genera, including Fusicatenibacter, Streptococcus, and Klebsiella, which may possess potential pathogenic properties. In particular, the bacterial genus Klebsiella demonstrated a positive correlation with irritable bowel syndrome severity scoring system scores. Conversely, healthy subjects showed enrichment of bacterial genera such as Alistipes, Akkermansia, and Dialister, which may be beneficial bacteria in IBS-D. Utilizing the random forest model, we developed a discriminative model for IBS-D based on differential bacterial genera. This model exhibited impressive performance, with an area under the curve value of 0.90. Additionally, our analysis did not reveal any gender-specific differences in the microbiota community composition among IBS-D patients. Conclusions Our findings offer preliminary insights into the potential relationship between intestinal microbiota and IBS-D. The identification model for IBS-D, grounded in gut microbiota, holds promising prospects for improving early diagnosis of IBS-D.

Alterations of brain structural and functional connectivity networks topology and decoupling in pediatric myelin oligodendrocyte glycoprotein antibody-associated disease

ObjectiveTo investigate the alteration in structural and functional connectivity networks (SCN and FCN) as well as their coupling in pediatric myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), and determine if these properties could serve as potential biomarkers for the disease. Materials and methodsIn total of 32 children with MOGAD and 30 age- and sex-matched healthy controls (HC) were employed to construct the SCN and FCN, respectively. The graph-theoretical analyses of the global properties, node properties of the 90 brain nodes, and the structural-functional connectivity (SC-FC) coupling of the two networks were performed. The graph-theoretical properties that exhibited significant differences were analyzed using partial correlation analysis in conjunction with the clinical scales, including the expanded disability status scale (EDSS), modified Rankin scale (mRS), and pediatric cerebral performance category (PCPC) of the MOGAD group. Subsequently, a machine learning model was developed to discriminate between MOGAD and the HC group, aiming to explore the potential of these properties as biomarkers. ResultsThe SCN of the MOGAD group exhibited aberrant global properties, including an increased characteristic path length (Lp) and a decreased global efficiency (Eg), along with reduced nodal properties such as degree centrality (Dc), nodal efficiency (Ne), and local efficiency in multiple nodes. The FCN of the MOGAD group only exhibited decreased Dc, Ne, and betweenness centrality in two nodes of nodal properties. Besides, MOGAD showed a significant decrease in SC-FC coupling compared to the HC group. The analysis of partial correlation revealed significant correlations between several properties and the scales of EDSS and mRS in the MOGAD group. The machine learning method was used to extract six features and establish the model, achieving a classification accuracy of 82.3% for MOGAD. ConclusionsPediatric MOGAD showed a more pronounced impairment in the SCN along with decoupling of SC-FC. Both partial correlation analysis and discriminant modeling suggest that alterations in brain network properties have the potential as biomarkers for assessing brain damage in MOGAD.

Predicting pathological response to neoadjuvant immunotherapy plus chemotherapy in resectable non-small cell lung cancer through liquid biopsy.

e20073 Background: Neoadjuvant immunotherapy (ICI) plus chemotherapy (CT) provides a significant benefit compared to CT alone in resectable non-small cell lung cancer (NSCLC). However, further research is needed to identify the populations benefitting from this combined strategy. Methods: Patients (pts) diagnosed with stage II-III resectable NSCLC who opted for neoadjuvant ICI plus CT were enrolled. Plasma samples were collected before treatment initiation (T1) and before surgery (T2). A next-generation sequencing (NGS) platform spanning 769 cancer-related genes was utilized to detect circulating tumor DNA (ctDNA). The ctDNA concentration was presented as haploid genome equivalents (hGE) per mL of plasma. Additionally, low-pass whole-methylome sequencing was employed to explore three features of cell-free DNA, including methylated fragment ratio (MFR), fragment size index (FSI), and chromosomal aneuploidy of featured fragments (CAFF). Clinical and pathological data was obtained from medical records. Results: Out of 24 pts enrolled, 18 pts were evaluable on both platforms. The major pathological response (MPR) rate was 70.8% (17/24), and the pathological complete response (pCR) rate was 37.5% (9/24). Squamous cell carcinoma (SCC) showed a higher MPR rate than adenocarcinoma (P = 0.014). Pts with clinical T stage 2 showed a trend toward a higher MPR rate than those with T stage 3/4 (P = 0.069). No single molecular biomarker was found to be independently associated with MPR or pCR at T1. A linear discriminant model was constructed based on clinical T stage, MFR, and hGE at T1, which predicted both MPR and pCR with an accuracy of 83.3% (n = 18). In pts with SCC (n = 15), the accuracy of this model in predicting MPR and pCR was 80.0% and 86.7%, respectively. Moreover, the CAFF score at T2 was significantly higher in non-MPR pts than in MPR pts (P = 0.003). The dynamic change of CAFF score [calculation: (T2-T1) / T1] was also significantly related to MPR (P = 0.006). Conclusions: Baseline cfDNA features and clinical parameters effectively predicted pathological response of neoadjuvant ICI plus CT in resectable NSCLC, which could potentially serve as a tool for identifying patients that would benefit most from the combined approach.

Unveiling the metabolic and coagulation disruptions in SARS-CoV-2-associated acute macular neuroretinopathy: A case-control study.

SARS-CoV-2 infection has been associated with the increased incidence of acute macular neuroretinopathy (AMN), an infrequent ocular disorder. However, the precise mechanisms underpinning AMN in the context of SARS-CoV-2 infection (AMN-SARS-CoV-2) remain elusive. In this case-control study, 14 patients diagnosed with AMN-SARS-CoV-2 between 2022/12 and 2023/3 were enrolled and compared with 14 SARS-CoV-2-infected individuals without AMN, who served as controls (SARS-CoV-2-no AMN). Metabolomic profiling using ultrahigh-performance liquid chromatography-online electrospray mass spectrometryrevealed significant alterations in serum metabolites in AMN-SARS-CoV-2 patients. Coagulation abnormalities were observed in AMN-SARS-CoV-2 patients, and their relationship with metabolic disorders was studied. Finally, a predictive model for AMN-SARS-CoV-2 was established. Seventy-sixupregulated and 42 downregulated metabolites were identified in AMN-SARS-CoV-2 cases. Notably, arginine metabolism within the urea cycle was significantly altered, evidenced by variations in ornithine, citrulline, l-proline, and ADAM levels, correlating with abnormal coagulation markers like platelet crit, fibrinogen degradation product, and fibrinogen. Additionally, increased arginase 1 (AGR1) activity within the urea cycle and reduced nitric oxide synthaseactivity were observed in AMN-SARS-CoV-2. The integration of urea cycle metabolite levels with coagulation parameters yielded a robust discriminatory model for AMN-SARS-CoV-2, as evidenced by an area under the curveof 0.96. The findings of the present study enhance our comprehension of the underlying metabolic mechanisms associated with AMN-SARS-CoV-2 and offer potential diagnostic markers for this uncommon ocular disorder within the context of SARS-CoV-2 infection.

Identification of mine water source based on TPE-LightGBM

Mine water inrush is a serious threat to mine safety production. It is very important to identify water inrush source types quickly to prevent and control water damage. In this study, the aqueous chemical components Na+ + K+, Ca2+, Mg2+, Cl−, SO42− and HCO3− of different aquifers in Pingdingshan coalfield were selected as the characteristic values, and the Surface water, Quaternary pore water, Carboniferous limestone karst water, Permian sandstone water, and Cambrian limestone karst water were used as the labels. An intelligent water source discrimination model is proposed by combining data mining, classification models, and reinforcement learning. As outlier data in the samples may interfere with the model recognition ability, the data distribution range was analyzed using box plots, and 20 groups of abnormal samples were excluded. The processed water chemistry data were divided into 80% learning samples and 20% test samples, and the learning samples were fed into a light gradient boosting machine (LightGBM) for training. The tree-structured parson estimator (TPE) obtains the optimal values of the main parameters of LightGBM in a very short time. Substituting the hyperparameters back into the model yields a 13.9% improvement in the accuracy of the model, proving the effectiveness of the TPE algorithm. To further validate the performance of the model, TPE-LightGBM is compared and analyzed with a Random Search-Multi Layer Perceptron Machine (RS-MLP) and Genetic Algorithm-Extreme Gradient Boosting Tree (GA-SVM). The accuracy of TPE-LightGBM, RS-MLP, and GA-SVM is 0.931, 0.759, 0.724 in that order, and the generalization error RMSE is 0.415, 1.05, and 1.313 in that order. The results show that TPE-LightGBM is more advantageous in water source identification and is more resistant to overfitting. By calculating and comparing the information gain of each variable, the contribution of Ca2+ is the highest, so it is necessary to pay attention to the change in Ca2+ concentration. TPE-LightGBM’s high accuracy and generalization ability have a good prospect for the identification of sudden water source types.

Machine Learning to Predict Pregnancy in Dairy Cows: An Approach Integrating Automated Activity Monitoring and On-Farm Data.

Automated activity monitoring (AAM) systems are critical in the dairy industry for detecting estrus and optimizing the timing of artificial insemination (AI), thus enhancing pregnancy success rates in cows. This study developed a predictive model to improve pregnancy success by integrating AAM data with cow-specific and environmental factors. Utilizing data from 1,054 cows, this study compared the pregnancy outcomes between two AI timings-8 or 10 h post-AAM alarm. Variables such as age, parity, body condition, locomotion, and vaginal discharge scores, peripartum diseases, the breeding program, the bull used for AI, milk production at the time of AI, and environmental conditions (season, relative humidity, and temperature-humidity index) were considered alongside the AAM data on rumination, activity, and estrus intensity. Six predictive models were assessed to determine their efficacy in predicting pregnancy success: logistic regression, Bagged AdaBoost algorithm, linear discriminant, random forest, support vector machine, and Bagged Classification Tree. Integrating the on-farm data with AAM significantly enhanced the pregnancy prediction accuracy at AI compared to using AAM data alone. The random forest models showed a superior performance, with the highest Kappa statistic and lowest false positive rates. The linear discriminant and logistic regression models demonstrated the best accuracy, minimal false negatives, and the highest area under the curve. These findings suggest that combining on-farm and AAM data can significantly improve reproductive management in the dairy industry.

Estimation of the earlywood and latewood ratio with different particle size of dry and wet milled Japanese cedar wood flour by using hyperspectral imaging

Mechanical properties of wood–plastic composites are influenced by a particle size and surface morphology of wood flour. Generally various sizes of wood flour are produced from single solid wood even if the single process is used. If the different particle size of wood flour is produced from different wood tissue such as earlywood (EW) and latewood (LW), not only particle size but also density and chemical composition of wood flour might influence the mechanical property of final products. This study aims to investigate the relationship between particle size and their origin; EW and LW. EW and LW were separately milled to produce the EW and LW flour by dry and wet milling. Hyperspectral images ranging 400–1000 nm for each wood flour were used as training data. Discriminant model of EW and LW flour developed by PLS-DA showed over 0.77 of accuracy. Then the EW and LW were dry and wet milled together and screened by three different sieve openings to obtain different particle size wood flour. Discriminant model was applied for the hyperspectral images of each size of wood flour to estimate the EW and LW ratio. The result showed that increasing sieve opening resulted in the increasing ratio of LW for dry milled wood flours. The results suggest that the EW was easily pulverized than LW.

Exploring Alternative Approaches to Language Modeling for Learning from Data and Knowledge

Despite their extensive application in language understanding tasks, large language models (LLMs) still encounter challenges including hallucinations - occasional fabrication of information - and alignment issues - lack of associations with human-curated world models (e.g., intuitive physics or common-sense knowledge). Moreover, the black-box nature of LLMs presents significant obstacles in training them effectively to achieve desired behaviors. In particular, modifying the concept embedding spaces of LLMs can be highly intractable. This process involves analyzing the implicit impact of such adjustments on the myriad parameters within LLMs and the resulting inductive biases. We propose a novel architecture that wraps powerful function approximation architectures within an outer, interpretable read-out layer. This read-out layer can be scrutinized to explicitly observe the effects of concept modeling during the training of the LLM. Our method stands in contrast with gradient-based implicit mechanisms, which depend solely on adjustments to the LLM parameters and thus evade scrutiny. By conducting extensive experiments across both generative and discriminative language modeling tasks, we evaluate the capabilities of our proposed architecture relative to state-of-the-art LLMs of similar sizes. Additionally, we offer a qualitative examination of the interpretable read-out layer and visualize the concepts it captures. The results demonstrate the potential of our approach for effectively controlling LLM hallucinations and enhancing the alignment with human expectations.

Comparative Analysis between Quality of Life and Human Labor in Countries Belonging to G7 and BRICS Blocks: Proposition of Discriminant Analysis Model

The aim of the present research is to identify and analyze the variables which help to effectively differentiate Quality of Life (QoL) and human labor in the G7 (Germany, France, Italy, Canada, Japan, United Kingdom, and United States of America—USA) and BRICS countries (Brazil, Russia, India, China, and South Africa) through a discriminant analysis. A discriminant analysis model is developed to classify countries as having a low, mid, or high QoL based on QoL and human labor variables. The variables used in the discriminant analysis were obtained between 2010 and 2022 from two platforms: NUMBEO variables capable of relating QoL to socioeconomic aspects and OECD’s (Organization for Economic Cooperation and Development) human-labor-related variables. Based on the results, the three variables that most discriminate the groups in order of importance are employed women in relation to the female population, the female labor force participation rate, and the female unemployment rate. Countries are classified as having a low, mid, or high QoL. The adopted technique will allow researchers and managers to classify and draw goals for action reorganization and investment in QoL and labor.

Physicochemical parameters combined flash GC e-nose and NIR for quality and volatile characterization of Zhenjiang vinegar with different aging time

In order to clarify the scientific connotation of the processing adjuvant vinegar of “the aging the better”, the physicochemical indexes, antioxidant activity in vitro, odor characteristics and NIR spectrum characteristics of Zhenjiang vinegar with different aging time were comprehensively studied, and the compositional characterization by UPLC-MS. The results showed that the physicochemical quality of Zhenjiang vinegar increased rapidly during the 2–3 years, entered a rising stable period during the 5–8 years and entered a sudden increase period during the 8–10 years of fermentation. Overall, it was in line with the trend that “the aging the better”. The study further characterized the odor changes of different aging time samples, established a fast and non-damage SVM discriminant model based on NIR spectral characteristics (accuracy 100 %), and 753 components were characterized by UPLC-MS, in order to provide a reference for the quality evaluation and standard improvement of Zhenjiang vinegar and processing adjuvant.

Associations between serum ferritin baselines and trajectories and the incidence of metabolic dysfunction-associated steatotic liver disease: a prospective cohort study.

Evidence from prospective cohort studies on the relationship between metabolic dysfunction-associated steatotic liver disease (MASLD) and longitudinal changes in serum ferritin (SF) still limited. This study aimed to investigate the associations of SF baselines and trajectories with new-onset MASLD and to present a MASLD discriminant model. A total of 1895 participants who attended health examinations at least three times in a hospital in Dalian City between 2015 and 2022 were included. The main outcome was the incidence of MASLD. The associations between SF baselines and trajectories with the risk of MASLD were analyzed by Cox proportional hazards regression, restricted cubic spline (RCS) analysis and time-dependent receiver operating characteristic (ROC) curve analysis. In addition, a MASLD discrimination model was established using logistic regression analyses. Among the 1895 participants, 492 developed MASLD during follow-up. Kaplan-Meier analysis indicated that participants in the low-stable trajectory group had a longer MASLD-free time compared with participants in other groups. Compared with those in the low-stable trajectory group, the adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the risk of new-onset MASLD in the medium-high, high-stable and high-high trajectory groups were 1.54(1.18-2.00), 1.77(1.35-2.32) and 1.55(1.07-2.26), respectively (Ptrend < 0.001). The results were robust in subgroup and sensitivity analyses. Multivariate Cox proportional regression showed that SF was an independent risk factor of MASLD (HR = 1.002, 95%CI: 1.000-1.003, P = 0.003). The restricted cubic spline demonstrated a nonlinear relationship between SF and the risk of MASLD. The 8-variable model had high discriminative performance, good accuracy and clinical effectiveness. The ROC curve results showed that AUC was greater than that of the FLI, HSI and ZJU models (all P < 0.01). Not only a higher baseline SF but also SF changing trajectory are significantly associated with risk of new-onset MASLD. SF could be a predictor of the occurrence of MASLD.

Stellar atmospheric parameters from Gaia BP/RP spectra using uncertain neural networks

ABSTRACT With the plentiful information available in the Gaia BP/RP spectra, there is significant scope for applying discriminative models to extract stellar atmospheric parameters and abundances. We describe an approach to leverage an ‘Uncertain Neural Network’ model trained on APOGEE data to provide high-quality predictions with robust estimates for per-prediction uncertainty. We report median formal uncertainties of 0.068 dex, 69.1 K, 0.14 dex, 0.031 dex, 0.040 dex, and 0.029 dex for [Fe/H], Teff, log g, [C/Fe], [N/Fe], and [α/M], respectively. We validate these predictions against our APOGEE training data, LAMOST, and Gaia GSP-phot stellar parameters, and see a strong correlation between our predicted parameters and those derived from these surveys. We investigate the information content of the spectra by considering the ‘attention’ our model pays to different spectral features compared to expectations from synthetic spectra calculations. Our model’s predictions are applied to the Gaia data set, and we produce a publicly available catalogue of our model’s predictions.

Mode Collapse Detection Strategies in Generative Adversarial Networks for Credit Card Fraud Detection

A Generative Adversarial Network (GAN) is an artifi- cial intelligence model developed specifically to pro- duce synthetic data that resembles real data by training a generative model and a discriminative model simulta- neously using adversarial training. A GAN can be ex- tensively used for generating replicated data, however, it suffers from several issues, one of which is mode col- lapse. Mode collapse takes place when the generator is unable to capture the complete range of diversity in the target data distribution, resulting in the production of limited and repeating variations of samples. Multiple metrics exist to quantify mode collapse in GANs, al- though no individual metric is capable of consistently providing accurate results. This research focuses on the critical need for accurate mode collapse detection tech- niques in GANs, to strengthen the credit card fraud detection systems. In this work, we utilize a GAN to generate numerical data instead of image data. Our ap- proach utilizes a wide range of measures, such as Gener- ator and Discriminator Loss, Wasserstein Distance, pre- cision, recall, and visualization tools, to provide a com- prehensive framework for detecting mode collapse. In addition, we introduce an alert mechanism that identi- fies possible mode collapse at an early stage, allowing for earlier intervention and modifications to the training process. We have further proposed suggestions regard- ing monitoring and analyzing generator and discrimina- tor loss values to identify potential instances of mode collapse to help the developer optimize GAN training and improve the quality of synthetic data.

Lattice-Based Generation of Euclidean Geometry Figures

We present a user-guided method to generate geometry figures appropriate for high school Euclidean geometry courses: a useful starting point for an intelligent tutoring system to provide meaningful, realistic figures for study. We first establish that a two-dimensional geometry figure can be represented abstractly using a complete, lattice we call a geometry figure lattice (GFL). As input, we take a user-defined vector of primitive geometry shapes and convert each into a GFL. We then exhaustively combine each these ‘primitive’ GFLs into a set of complex GFLs using a process we call gluing. We mitigate redundancy in GFLs by introducing a polynomial-time algorithm for determining if two GFLs are isomorphic. These lattices act as a template for the second step: instantiating GFLs into a sequence of concrete geometry figures. To identify figures that are structurally similar to textbook problems, we use a discriminator model trained on a corpus of textbook geometry figures.