Data Structure Research Articles

Divalent Titanium via Reductive N−C Coupling of a TiIV Nitrido with π‐Acids

AbstractThe nitrido‐ate complex [(PN)2Ti(N){μ2‐K(OEt2)}]2 (1) (PN−=(N‐(2‐PiPr2‐4‐methylphenyl)‐2,4,6‐Me3C6H2) reductively couples CO and isocyanides in the presence of DME or cryptand (Kryptofix222), to form rare, five‐coordinate TiII complexes having a linear cumulene motif, [K(L)][(PN)2Ti(NCE)] (E=O, L=Kryptofix222, (2); E=NAd, L=3 DME, (3); E=NtBu, L=3 DME, (4); E=NAd, L=Kryptofix222, (5)). Oxidation of 2–5 with [Fc][OTf] afforded an isostructural TiIII center containing a neutral cumulene, [(PN)2Ti(NCE)] (E=O, (6); E=NAd (7), NtBu (8)) and characterization by CW X‐band EPR spectroscopy, revealed unpaired electron to be metal centric. Moreover, 1e− reduction of 6 and 7 in the presence of Kryptofix222cleanly reformed corresponding discrete TiII complexes 2 and 5, which were further characterized by solution magnetization measurements and high‐frequency and ‐field EPR (HFEPR) spectroscopy. Furthermore, oxidation of 7 with [Fc*][B(C6F5)4] resulted in a ligand disproportionated TiIV complex having transoid carbodiimides, [(PN)2Ti(NCNAd)2] (9). Comparison of spectroscopic, structural, and computational data for the divalent, trivalent, and tetravalent systems, including their 15N enriched isotopomers demonstrate these cumulenes to decrease in order of backbonding as TiII→TiIII→TiIV and increasing order of π‐donation as TiII→TiIII→TiIV, thus displaying more covalency in TiIII species. Lastly, we show a synthetic cycle whereby complex 1 can deliver an N‐atom to CO and CNAd.

XGBoost-Enhanced Graph Neural Networks: A New Architecture for Heterogeneous Tabular Data

Graph neural networks (GNNs) perform well in text analysis tasks. Their unique structure allows them to capture complex patterns and dependencies in text, making them ideal for processing natural language tasks. At the same time, XGBoost (version 1.6.2.) outperforms other machine learning methods on heterogeneous tabular data. However, traditional graph neural networks mainly study isomorphic and sparse data features. Therefore, when dealing with tabular data, traditional graph neural networks encounter challenges such as data structure mismatch, feature selection, and processing difficulties. To solve these problems, we propose a novel architecture, XGNN, which combines the advantages of XGBoost and GNNs to deal with heterogeneous features and graph structures. In this paper, we use GAT for our graph neural network model. We can train XGBoost and GNN end-to-end to fit and adjust the new tree in XGBoost based on the gradient information from the GNN. Extensive experiments on node prediction and node classification tasks demonstrate that the performance of our proposed new model is significantly improved for both prediction and classification tasks and performs particularly well on heterogeneous tabular data.

Implementing Linked Art in a Multi-Modal Database for Cross-Collection Discovery

Yale University has implemented a knowledge graph based discovery system that brings together the various art, natural history, archival, conservation and bibliographic collections using Linked Open Usable Data standards such as Linked Art and IIIF. This system comprises more than 41 million records, which would expand to more than 2 billion RDF triples, and is thus at a similar scale to Europeana. This paper presents the lessons learned from the five year effort around the usability of linked data structures across the organization, the technologies needed to make use of the knowledge in a performant way, and the appropriate design paradigms for front end applications which make the graph easily and intuitively accessible to researchers and the public, including the necessity of consistency in data modeling, that records are an essential concept to maintain through multi-modal systems, and the use of hypertext and web caches to maintain the separation between systems.

P-754 Association between ovarian hyperstimulation syndrome and adverse obstetric and neonatal outcomes in Korea: a nationwide health insurance database study

Abstract Study question Does OHSS affect maternal and neonatal outcomes in perinatal life? Summary answer OHSS is associated with an increased risk of postpartum hemorrhage, large birth weight, neonatal respiratory distress syndrome, and intraventricular hemorrhage. What is known already Ovarian hyperstimulation syndrome (OHSS) is an iatrogenic condition caused by ovulation induction. Golan et al.(1989)’s clinical staging system categorizes OHSS into mild, moderate, and severe based on the clinical characteristics. Among them, severe OHSS is considered potentially fatal and requires hospitalization. In 2021, a systematic review of 3,303 pregnant patients with OHSS and 89,720 pregnant patients without OHSS, drawn from 13 studies, revealed a notable impact of OHSS on preterm birth. However, owing to the characteristics of spontaneous resolution in the relatively early stages of pregnancy, no large-scale study has confirmed the OHSS and pregnancy outcomes. Study design, size, duration This retrospective study utilized a combined dataset from two primary sources: the Korea National Health Insurance (KNHI) claims database and the National Health Screening Program for Infants and Children (NHSP-IC) database. We identified all women who delivered between January 1, 2012, and December 31, 2020 from KNHI claims database. Pregnancy outcomes including preterm birth and PPH were identified from NHSP-IC database and KNHI claims database, respectively. Participants/materials, setting, methods We enrolled all women who delivered between January 1, 2012, and December 31, 2020 (N = 3,446,520). Women diagnosed with multiple pregnancy or had missing data were excluded (N = 2,565,461). The participants were divided into three groups according to the presence or absence of an ovarian hyperstimulation diagnosis code (ICD-10 code N981) and hospitalization within 3 months before pregnancy: non-OHSS group (N = 2,562,562), mild-to-moderate OHSS group (N = 2,293), severe OHSS group (N = 606). Main results and the role of chance In this study, mothers in the ovarian hyperstimulation group had a younger age and a higher rate incidence of primiparity and cesarean section. In the ovarian hyperstimulation group, a significantly high proportion of patients was diagnosed with pre-pregnancy hypertension or diabetes mellitus and gestational hypertension or gestational diabetes. The risk of postpartum hemorrhage was 38.2% higher in the mild-to-moderate OHSS group than in the non-OHSS group (adjusted odds ratio (aOR): 1.382, 95% confidence interval (CI): 1.223–1.562). No difference was found in the average birth weight of newborns according to the presence or absence of OHSS. However, the risk of macrosomia was elevated by 62.4% in the severe OHSS group (aOR: 1.624, 95% CI: 1.168–2.259) than the non-OHSS group. The mild-to-moderate OHSS group had a 36.8% higher risk of respiratory distress syndrome compared with the non-OHSS group (aOR: 1.368, 95% CI: 1.149–1.629). The mild-to-moderate OHSS group had a 2.697-fold increased risk of intraventricular hemorrhage compared with the non-OHSS group (aOR: 2.697, 95% CI: 1.338–5.437). Limitations, reasons for caution A limitation lies in the unknown details of the controlled ovarian stimulation protocol, type and dose of gonadotrophin used, hormone concentration, and embryo transfer method used due to the data structure. Despite the potential variations in the causes of hospitalization, hospitalized patients were classified into the severe OHSS group. Wider implications of the findings Although OHSS occurs during early pregnancy, further research is required to determine its long-term effects. Especially, if the incidence of IVH is significantly higher in the OHSS group, particularly among uncommon full-term neonates, further studies are warranted. Trial registration number Not Applicable

Brazing of Porous Nickel to Copper and Stainless Steel: Microstructure and Mechanical Properties

A porous nickel (Ni) was brazed to copper (Cu) and stainless steel 304 (SS304) using VZ2250 and MBF67 brazing filler metal with a composition of 77.4Cu-9.3Sn-7.0Ni-6.3P and 64.5Ni-25Cr-6P-1.5Si (Cu: Copper, Sn: Tin, Ni: Nickel, P: Phosphorus, Cr: Chromium, Si: Silicon), respectively for joint microstructure and mechanical properties analysis. Porous Ni with a pore density of 15 pores per inch (PPI) was sandwiched between Cu/VZ2250 and MBF67/SS304. A brazed joint of Cu/Porous Ni/SS304 with VZ2250 and MBF67 brazing filler metal was prepared in a high vacuum furnace at different brazing times of 5, 10, and 15 minutes for 1015 °C with a heating and cooling rate of 10 °C/min, respectively for comparison purpose. The microstructure and mechanical properties of the brazed joint were investigated to identify the joint ability after the brazing process. Scanning Electron Microscope (SEM) equipped with Energy Dispersive X-Ray Spectroscopy (EDS) confirmed the interfacial microstructure by the formation of the diffusion filler metal (dark grey colour) for the Cu/Porous Ni/SS304 with VZ2250 and MBF67 brazing filler metal. For shear strength tests, the value decreases with an increase in the brazing time. The shear strength tests for the brazed joint of Cu/Porous Ni/SS304 with VZ2250 and MBF67 brazing filler metal show the maximum shear strength test value can be achieved for the brazing time of 5 minutes. The decreasing shear strength value was observed with differences in structural data of porous Ni due to the softening after the brazing process. Keywords: Brazing, Microstructure, Porous Nickel, Shear Strength.

Causal effects of hypertensive disorders of pregnancy on structural changes in specific brain regions: a Mendelian randomization study.

This study utilized Mendelian randomization to explore the impact of hypertensive disorders of pregnancy and their subtypes on brain structures, using genome-wide association study data from the FinnGen consortium for hypertensive disorders of pregnancy exposure and brain structure data from the ENIGMA consortium as outcomes. The inverse-variance weighted method, along with Cochran's Q test, Mendelian randomization-Egger regression, Mendelian randomization-PRESSO global test, and the leave-one-out approach, were applied to infer causality and assess heterogeneity and pleiotropy. Findings indicate hypertensive disorders of pregnancy are associated with structural brain alterations, including reduced cortical thickness in areas like the insula, isthmus cingulate gyrus, superior temporal gyrus, temporal pole, and transverse temporal gyrus, and an increased surface area in the superior frontal gyrus. Specific associations were found for hypertensive disorders of pregnancy subtypes: chronic hypertension with superimposed preeclampsia increased cortical thickness in the supramarginal gyrus; preeclampsia/eclampsia led to thinner cortex in the lingual gyrus and larger hippocampal volume and superior parietal lobule surface area. Chronic hypertension was associated with reduced cortical thickness in the caudal and rostral anterior cingulate and increased surface area of the cuneus and thickness of the pars orbitalis cortex. Gestational hypertension showed no significant brain region changes. These insights clarify hypertensive disorders of pregnancies' neurological and cognitive effects by identifying affected brain regions.

Semantic clustering method using integration of advanced LDA algorithm and BERT algorithm

The subject of the study is an in-depth semantic data analysis based on the modification of the Latent Dirichlet Allocation (LDA) methodology and its integration with the bidirectional encoding representation of transformers (BERT). Relevance. Latent Dirichlet Allocation (LDA) is a fundamental topic modeling technique that is widely used in a variety of text analysis applications. Although its usefulness is widely recognized, traditional LDA models often face limitations, such as a rigid distribution of topics and inadequate representation of semantic nuances inherent in natural language. The purpose and main idea of the study is to improve the adequacy and accuracy of semantic analysis by improving the basic LDA mechanism that integrates adaptive Dirichlet priorities and exploits the deep semantic capabilities of BERT embeddings. Research methods: 1) selection of textual datasets; 2) data preprocessing steps; 3) improvement of the LDA algorithm; 4) integration with BERT Embeddings; 5) comparative analysis. Research objectives: 1) theoretical substantiation of LDA modification; 2) implementation of integration with BERT; 3) evaluation of the method efficiency; 4) comparative analysis; 5) development of an architectural solution. The results of the research are that, first of all, the theoretical foundations of both the standard and modified LDA models are outlined, and their extended formula is presented in detail. Through a series of experiments on text datasets characterized by different emotional states, we emphasize the key advantages of the proposed approach. Based on a comparative analysis of such indicators as intra- and inter-cluster distances and silhouette coefficient, we prove the increased coherence, interpretability, and adaptability of the modified LDA model. An architectural solution for implementing the method is proposed. Conclusions. The empirical results indicate a significant improvement in the detection of subtle complexities and thematic structures in textual data, which is a step in the evolutionary development of thematic modeling methodologies. In addition, the results of the research not only open up the possibility of applying LDA to more complex linguistic scenarios, but also outline ways to further improve them for unsupervised text analysis.

Proximity Graph Networks: Predicting Ligand Affinity with Message Passing Neural Networks.

Message passing neural networks (MPNNs) on molecular graphs generate continuous and differentiable encodings of small molecules with state-of-the-art performance on protein-ligand complex scoring tasks. Here, we describe the proximity graph network (PGN) package, an open-source toolkit that constructs ligand-receptor graphs based on atom proximity and allows users to rapidly apply and evaluate MPNN architectures for a broad range of tasks. We demonstrate the utility of PGN by introducing benchmarks for affinity and docking score prediction tasks. Graph networks generalize better than fingerprint-based models and perform strongly for the docking score prediction task. Overall, MPNNs with proximity graph data structures augment the prediction of ligand-receptor complex properties when ligand-receptor data are available.

The method of linear-logical operators and logical equations in information extraction tasks

Relational and logical methods of knowledge representation play a key role in creating a mathematical basis for information systems. Predicate algebra and predicate operators are among the most effective tools for describing information in detail. These tools make it easy to formulate formalized information, create database queries, and simulate human activity. In the context of the new need for reliable and efficient data selection, a problem arises in deeper analysis. Subject of the study is the theory of quantum linear equations based on the algebra of linear predicate operations, the formal apparatus of linear logic operators and methods for solving logical equations in information extraction tasks. Aim of the study is a developing of a method for using linear logic operators and logical equations to extract information. This approach can significantly optimize the process of extracting the necessary information, even in huge databases. Main tasks: analysis of existing approaches to information extraction; consideration of the theory of linear logic operators; study of methods for reducing logic to an algebraic form; analysis of logical spaces and the algebra of finite predicate actions and the theory of linear logic operators. The research methods involve a systematic analysis of the mathematical structure of the algebra of finite predicates and predicate functions to identify the key elements that affect the query formation process. The method of using linear logic operators and logical equations for information extraction is proposed. The results of the study showed that the method of using linear logic operators and logical equations is a universal and adaptive tool for working with algebraic data structures. It can be applied in a wide range of information extraction tasks and prove its value as one of the possible methods of information processing. Conclusion. The paper investigates formal methods of intelligent systems, in particular, ways of representing knowledge in accordance with the peculiarities of the field of application and the language that allows encoding this knowledge for storage in computer memory. The proposed method can be implemented in the development of language interfaces for automated information access systems, in search engine algorithms, for logical analysis of information in databases and expert systems, as well as in performing tasks related to object recognition and classification.

Semantic Visual SLAM Algorithm Based on Improved DeepLabV3+ Model and LK Optical Flow

Aiming at the problem that dynamic targets in indoor environments lead to low accuracy and large errors in the localization and position estimation of visual SLAM systems and the inability to build maps containing semantic information, a semantic visual SLAM algorithm based on the semantic segmentation network DeepLabV3+ and LK optical flow is proposed based on the ORB-SLAM2 system. First, the dynamic target feature points are detected and rejected based on the lightweight semantic segmentation network DeepLabV3+ and LK optical flow method. Second, the static environment occluded by the dynamic target is repaired using the time-weighted multi-frame fusion background repair technique. Lastly, the filtered static feature points are used for feature matching and position calculation. Meanwhile, the semantic labeling information of static objects obtained based on the lightweight semantic segmentation network DeepLabV3+ is fused with the static environment information after background repair to generate dense point cloud maps containing semantic information, and the semantic dense point cloud maps are transformed into semantic octree maps using the octree spatial segmentation data structure. The localization accuracy of the visual SLAM system and the construction of the semantic maps are verified using the widely used TUM RGB-D dataset and real scene data, respectively. The experimental results show that the proposed semantic visual SLAM algorithm can effectively reduce the influence of dynamic targets on the system, and compared with other advanced algorithms, such as DynaSLAM, it has the highest performance in indoor dynamic environments in terms of localization accuracy and time consumption. In addition, semantic maps can be constructed so that the robot can better understand and adapt to the indoor dynamic environment.

Deal.t: an implementation of multivariate analysis suitable T-splines within the deal.ii framework

AbstractWe present a numerical framework for solving partial differential equations within an isogeometric context using T-splines in two and three space dimensions. Within this paper, we explain the data structures used for the implementation of ( with T-splines) and main differences when using in contrast to . The authors present numerical experiments with error-based refinement (2D) and a priori refinement (3D) for scalar-valued problems. A full tutorial is given in the appendix. Since the new framework is based on , T-splines may be applied to various different PDEs.

Accurate Prediction of Protein Structural Flexibility by Deep Learning Integrating Intricate Atomic Structures and Cryo-EM Density Information

The dynamics of proteins are crucial for understanding their mechanisms. However, computationally predicting protein dynamic information has proven challenging. Here, we propose a neural network model, RMSF-net, which outperforms previous methods and produces the best results in a large-scale protein dynamics dataset; this model can accurately infer the dynamic information of a protein in only a few seconds. By learning effectively from experimental protein structure data and cryo-electron microscopy (cryo-EM) data integration, our approach is able to accurately identify the interactive bidirectional constraints and supervision between cryo-EM maps and PDB models in maximizing the dynamic prediction efficacy. Rigorous 5-fold cross-validation on the dataset demonstrates that RMSF-net achieves test correlation coefficients of 0.746 ± 0.127 at the voxel level and 0.765 ± 0.109 at the residue level, showcasing its ability to deliver dynamic predictions closely approximating molecular dynamics simulations. Additionally, it offers real-time dynamic inference with minimal storage overhead on the order of megabytes. RMSF-net is a freely accessible tool and is anticipated to play an essential role in the study of protein dynamics.

Fabrication of CdS NRs /MoS2@g-C3N4 nanotubes for efficient photoelectrochemical hydrogen production

Exploration of noble metal-free and highly efficient photoelectrocatalyst for hydrogen production attracts the attention of researchers. The graphitic carbon nitride nanotubes, cadmium sulfide nanorods (NRs) and molybedun disulfide, CdS NRs/MoS2@g-C3N4 nanotubes were synthesized via facile hydrothermal method for hydrogen production studies. The synthesized CdS NRs/MoS2@g-C3N4 nanotubes were studied for electrochemical hydrogen production applications. Structural data revealed that the synthesized composite materials attained crystallinity. Electrochemical investigations demonstrated the highest hydrogen production efficiency of CdS NRs/MoS2@g-C3N4 nanotubes as compared to the bare graphitic carbon nitride nanotubes. This seems that the modification of g-C3N4 NTs with MoS2 increased the photocatalytic efficiency of g-C3N4 NTs. The good efficiency of the CdS NRs/MoS2@g-C3N4 NTs is owing to a unique tubular structure, the presence of active sites, light reflection channel, nanotube composite heterostructure, electron-hole separation, better charge transfer and the synergistic effect. Moreover, it was observed that the tubular structure would be the best photoelectrocatalyst for hydrogen production achieving a geometric current density of 10 mAcm−2 at an overpotential of 0.81 V gives 144 mVdec−1 in 1 M KOH versus the RHE. These results may provide insights into constructing highly efficient photoelectrocatalysts by varying the structural morphology and heterojunction for hydrogen production.

Assessment of Mycological Possibility Using Machine Learning Models for Effective Inclusion in Sustainable Forest Management

The integral role of wild fungi in ecosystems, including provisioning, regulating, cultural, and supporting services, is well recognized. However, quantifying and predicting wild mushroom yields is challenging due to spatial and temporal variability. In Mediterranean forests, climate-change-induced droughts further impact mushroom production. Fungal fruiting is influenced by factors such as climate, soil, topography, and forest structure. This study aims to quantify and predict the mycological potential of Lactarius deliciosus in sustainably managed Mediterranean pine forests using machine learning models. We utilize a long-term dataset of Lactarius deliciosus yields from 17 Pinus pinaster plots in Soria, Spain, integrating forest-derived structural data, NASA Landsat mission vegetation indices, and climatic data. The resulting multisource database facilitates the creation of a two-stage ‘mycological exploitability’ index, crucial for incorporating anticipated mycological production into sustainable forest management, in line with what is usually done for other uses such as timber or game. Various Machine Learning (ML) techniques, such as classification trees, random forest, linear and radial support vector machine, and neural networks, were employed to construct models for classification and prediction. The sample was always divided into training and validation sets (70-30%), while the differences were found in terms of Overall Accuracy (OA). Neural networks, incorporating critical variables like climatic data (precipitation in January and humidity in November), remote sensing indices (Enhanced Vegetation Index, Green Normalization Difference Vegetation Index), and structural forest variables (mean height, site index and basal area), produced the most accurate and unbiased models (OAtraining = 0.8398; OAvalidation = 0.7190). This research emphasizes the importance of considering a diverse array of ecosystem variables for quantifying wild mushroom yields and underscores the pivotal role of Artificial Intelligence (AI) tools and remotely sensed observations in modeling non-wood forest products. Integrating such models into sustainable forest management plans is crucial for recognizing the ecosystem services provided by them.

Exploring Bacillus species xylanases for industrial applications: screening via thermostability and reaction modelling.

Xylanases derived from Bacillus species hold significant importance in various large-scale production sectors, with increasing demand driven by biofuel production. However, despite their potential, the extreme environmental conditions often encountered in production settings have led to their underutilisation. To address this issue and enhance their efficacy under adverse conditions, we conducted a theoretical investigation on a group of five Bacillus species xylanases belonging to the glycoside hydrolase GH11 family. Bacillus sp. NCL 87-6-10 (sp_NCL 87-6-10) emerged as a potent candidate among the selected biocatalysts; this Bacillus strain exhibited high thermal stability and achieved a transition state with minimal energy requirements, thereby accelerating the biocatalytic reaction process. Our approach aims to provide support for experimentalists in the industrial sector, encouraging them to employ structural-based reaction modelling scrutinisation to predict the ability of targeted xylanases. Utilising crystal structure data available in the Carbohydrate-Active enzymes database, we aimed to analyse their structural capabilities in terms of thermal-stability and activity. Our investigation into identifying the most prominent Bacillus species xylanases unfolds with the help of the semi-empirical quantum mechanics MOPAC method integrated with the DRIVER program is used in calculations of reaction pathways to understand the activation energy. Additionally, we scrutinised the selected xylanases using various analyses, including constrained network analyses, intermolecular interactions of the enzyme-substrate complex and molecular orbital assessments calculated using the AM1 method with the MO-G model (MO-G AM1) to validate their reactivity.

Using electronic health records for clinical pharmacology research: Challenges and considerations.

Electronic health records (EHRs) contain a vast array of phenotypic data on large numbers of individuals, often collected over decades. Due to the wealth of information, EHR data have emerged as a powerful resource to make first discoveries and identify disparities in our healthcare system. While the number of EHR-based studies has exploded in recent years, most of these studies are directed at associations with disease rather than pharmacotherapeutic outcomes, such as drug response or adverse drug reactions. This is largely due to challenges specific to deriving drug-related phenotypes from the EHR. There is great potential for EHR-based discovery in clinical pharmacology research, and there is a critical need to address specific challenges related to accurate and reproducible derivation of drug-related phenotypes from the EHR. This review provides a detailed evaluation of challenges and considerations for deriving drug-related data from EHRs. We provide an examination of EHR-based computable phenotypes and discuss cutting-edge approaches to map medication information for clinical pharmacology research, including medication-based computable phenotypes and natural language processing. We also discuss additional considerations such as data structure, heterogeneity and missing data, rare phenotypes, and diversity within the EHR. By further understanding the complexities associated with conducting clinical pharmacology research using EHR-based data, investigators will be better equipped to design thoughtful studies with more reproducible results. Progress in utilizing EHRs for clinical pharmacology research should lead to significant advances in our ability to understand differential drug response and predict adverse drug reactions.

Late Devonian felsic magmatism in southern New Brunswick and its association with a large igneous province that may have contributed to the Frasnian-Famennian extinction

Felsic rocks of the Piskahegan Group and coeval plutons form a major part of a Late Paleozoic epicontinental caldera in southwestern New Brunswick, Canada. The caldera forms an elliptical structure with a length of ∼34 km and a width of ∼13 km. It hosts a significant polymetallic deposit of tin, molybdenum, indium and bismuth associated with mid-sequence granitic intrusions. The caldera formed in the aftermath of the late Early to Middle Devonian Acadian Orogeny during the opening of the late Paleozoic Maritimes Basin in the Northern Appalachian Belt. Its surrounding successions are largely composed of bimodal igneous rocks that were derived from two distinct sources: upper mantle-derived mafic magma and lower crust-derived felsic magma. The rocks that marked the beginning of volcanic activity in the caldera are no younger than 374.2 ± 2 Ma, whereas the uppermost dated volcanic unit in the complex yielded an age of 364.6 ± 0.7 Ma. The latter is conformably overlain by a succession of red beds and undated basalts that also overlie penecontemporaneous felsic volcanic rocks adjacent to the caldera. The post-orogenic felsic rocks are fractionated, peraluminous A-2 type rocks with silica contents ranging from ∼70 to ∼79 wt% along with high K2O contents and ɛNd(t) ranging from −0.10 to +1.05. They were generated by the melting of lower crustal basement rocks linked to a lithospheric plate that was metasomatized during the Neoproterozoic. Based on available geochemical, geophysical and structural data on Late Paleozoic rocks of southeastern Canada, the melting may have been triggered by the injection of profuse mafic magma that eventually formed a thick underplating at the base of the crust in association with heat derived from an underlying mantle plume and conveyed by transtensional structures. Basal rocks of the Piskahegan Group (the Intracaldera Sequence) and penecontemporaneous volcanic rocks in Nova Scotia are interpreted as small erosional remnants of a Large Igneous Province (LIP) based on the extensive exposure of large coeval plutons throughout southeastern Canada. The new geochronological data suggests that profuse late Frasnian magmatism in this LIP could have contributed to the environmental deterioration that led to a significant extinction at the Frasnian-Famennian boundary (the Kellwasser Event).

Leveraging GPT-4 for identifying cancer phenotypes in electronic health records: a performance comparison between GPT-4, GPT-3.5-turbo, Flan-T5, Llama-3-8B, and spaCy's rule-based and machine learning-based methods.

Accurately identifying clinical phenotypes from Electronic Health Records (EHRs) provides additional insights into patients' health, especially when such information is unavailable in structured data. This study evaluates the application of OpenAI's Generative Pre-trained Transformer (GPT)-4 model to identify clinical phenotypes from EHR text in non-small cell lung cancer (NSCLC) patients. The goal was to identify disease stages, treatments and progression utilizing GPT-4, and compare its performance against GPT-3.5-turbo, Flan-T5-xl, Flan-T5-xxl, Llama-3-8B, and 2 rule-based and machine learning-based methods, namely, scispaCy and medspaCy. Phenotypes such as initial cancer stage, initial treatment, evidence of cancer recurrence, and affected organs during recurrence were identified from 13646 clinical notes for 63 NSCLC patients from Washington University in St. Louis, Missouri. The performance of the GPT-4 model is evaluated against GPT-3.5-turbo, Flan-T5-xxl, Flan-T5-xl, Llama-3-8B, medspaCy, and scispaCy by comparing precision, recall, and micro-F1 scores. GPT-4 achieved higher F1 score, precision, and recall compared to Flan-T5-xl, Flan-T5-xxl, Llama-3-8B, medspaCy, and scispaCy's models. GPT-3.5-turbo performed similarly to that of GPT-4. GPT, Flan-T5, and Llama models were not constrained by explicit rule requirements for contextual pattern recognition. spaCy models relied on predefined patterns, leading to their suboptimal performance. GPT-4 improves clinical phenotype identification due to its robust pre-training and remarkable pattern recognition capability on the embedded tokens. It demonstrates data-driven effectiveness even with limited context in the input. While rule-based models remain useful for some tasks, GPT models offer improved contextual understanding of the text, and robust clinical phenotype extraction.

Synchronous high-speed measurements of a flexible structure under wind gust load

Simultaneously measuring the fluid flow around a flexible structure and the resulting deformations during short-term yet highly dynamic flow events is the focus of this fluid–structure interaction (FSI) study. These scenarios occur when a wind gust impacts a flexible structure, leading to extreme loads and significant deflections. To mimic such gusts, a specifically designed wind gust generator is used within a wind tunnel featuring an open test section. A high-speed particle-image velocimetry system records the flow field, while the digital-image correlation technique captures the structural deformations. That allows us to perform synchronized coupled fluid–structure measurements for a T-structure under wind gust load. The time-resolved measurements are repeated up to 104 times, allowing for phase-averaging of both the flow and the structural data, and to examine the convergence of the statistics. A comprehensive analysis of the instantaneous and phase-averaged data reveals that the flow field in the vicinity of the structure undergoes noticeable changes during the gust impact. The recirculation region behind the T-structures perceptibly increases when the gust hits the structure. A maximum deformation of about 10% of its height is observed during the highly dynamic gust event. Given (1) the availability of synchronously recorded data for both the fluid flow and the structure deformation, (2) the simplicity of the structure's geometry, and (3) the moderate Reynolds number of about 4×104, this case also serves as a well-suited benchmark test case for evaluating simulation methodologies for strongly coupled, highly dynamic FSI problems.

T-distributed Stochastic Neighbor Network for unsupervised representation learning

Unsupervised representation learning (URL) is still lack of a reasonable operator (e.g. convolution kernel) for exploring meaningful structural information from generic data including vector, image and tabular data. In this paper, we propose a simple end-to-end T-distributed Stochastic Neighbor Network (TsNet) for URL with clustering downstream task. Concretely, our TsNet model has three major components: (1) an adaptive connectivity distribution learning module is presented to construct a pairwise graph for preserving the local structure of generic data; (2) a T-distributed stochastic neighbor embedding based loss function is designed to learn a transformation between embeddings and original data, which improves the discrimination of representations; (3) a nonlinear parametric mapping is learned via our TsNet on an unsupervised generalized manner, which can address the “out-of-sample” issue. By combining these components, our method is able to considerably outperform previous related unsupervised learning approaches on visualization and clustering of generic data. A simple deep neural network equipped on our model respectively achieves 74.90%, 76.56% ACC and NMI, which is 8% relative improvement over previous state-of-the-art on real single-cell RNA-sequencing (scRNA-seq) datasets clustering.