Model Predictions Research Articles

Mental representation of words and concepts in late multilingualism

Abstract Speakers of multiple languages must store the respective lexical items efficiently to enable correct access. Importantly, all items must be linked to semantic information and world knowledge. One prominent model of the mental lexicon of late bilinguals is the Revised Hierarchical Model (Kroll & Stewart, 1994), which postulates bidirectional but asymmetrical connections between separate stores for L1 (native language) and L2 (second/foreign language), and a shared conceptual store. Using German native speakers with advanced English proficiency, Experiment 1 largely confirmed model predictions regarding different preferred mental routes and processing times depending on translation direction. Moreover, the original design was extended by including abstract stimuli and picture naming in L2. A series of additional measures, such as proficiency and age of acquisition, served to specify the language experience of the participants and made it possible to compare the results with a group of non-native speakers of German (Experiment 2). Interestingly, the results suggest that the model also applies to two or more non-native languages, potentially influenced by the experimental and environmental language context.

Machine-Learning Prediction of Curie Temperature from Chemical Compositions of Ferromagnetic Materials.

Room-temperature ferromagnets are high-value targets for discovery given the ease by which they could be embedded within magnetic devices. However, the multitude of potential interactions among magnetic ions and their surrounding environments renders the prediction of thermally stable magnetic properties challenging. Therefore, it is vital to explore methods that can effectively screen potential candidates to expedite the discovery of novel ferromagnetic materials within highly intricate feature spaces. To this end, we explore machine-learning (ML) methods as a means to predict the Curie temperature (Tc) of ferromagnetic materials by discerning patterns within materials databases. This study emphasizes the importance of feature analysis and selection in ML modeling and demonstrates the efficacy of our gradient-boosted statistical feature-selection workflow for training predictive models. The models are fine-tuned through Bayesian optimization, using features derived solely from the chemical compositions of the materials data, before the model predictions are evaluated against literature values. We have collated ca. 35,000 Tc values and the performance of our workflow is benchmarked against state-of-the-art algorithms, the results of which demonstrate that our methodology is superior to the majority of alternative methods. In a 10-fold cross-validation, our regression model realized an R2 of (0.92 ± 0.01), an MAE of (40.8 ± 1.9) K, and an RMSE of (80.0 ± 5.0) K. We demonstrate the utility of our ML model through case studies that forecast Tc values for rare-earth intermetallic compounds and generate magnetic phase diagrams for various chemical systems. These case studies highlight the importance of a systematic approach to feature analysis and selection in enhancing both the predictive capability and interpretability of ML models, while being devoid of potential human bias. They demonstrate the advantages of such an approach over a mere reliance on algorithmic complexity and a black-box treatment in ML-based modeling within the domain of computational materials science.

Development of machine learning models for forecasting the strength of resilient modulus of subgrade soil: genetic and artificial neural network approaches

Accurately predicting the Modulus of Resilience (MR) of subgrade soils, which exhibit non-linear stress–strain behaviors, is crucial for effective soil assessment. Traditional laboratory techniques for determining MR are often costly and time-consuming. This study explores the efficacy of Genetic Programming (GEP), Multi-Expression Programming (MEP), and Artificial Neural Networks (ANN) in forecasting MR using 2813 data records while considering six key parameters. Several Statistical assessments were utilized to evaluate model accuracy. The results indicate that the GEP model consistently outperforms MEP and ANN models, demonstrating the lowest error metrics and highest correlation indices (R2). During training, the GEP model achieved an R2 value of 0.996, surpassing the MEP (R2 = 0.97) and ANN (R2 = 0.95) models. Sensitivity and SHAP (SHapley Additive exPlanations) analysis were also performed to gain insights into input parameter significance. Sensitivity analysis revealed that confining stress (21.6%) and dry density (26.89%) are the most influential parameters in predicting MR. SHAP analysis corroborated these findings, highlighting the critical impact of these parameters on model predictions. This study underscores the reliability of GEP as a robust tool for precise MR prediction in subgrade soil applications, providing valuable insights into model performance and parameter significance across various machine-learning (ML) approaches.

Enhanced lake elevation mapping using a zone-based method

Abstract Inland lakes play a crucial role in monitoring global climate change and managing responses to extreme weather events, with lake elevation being critical for assessing their regulatory capacities. However, due to the limited temporal resolution of current altimetry satellites, obtaining high-frequency, high-precision elevation data for water bodies remains challenging. Consequently, most studies utilize elevation-area (E-A) models constructed from historical elevation and area records, integrated with area observations from high-temporal resolution optical satellites to infer precise water levels. Yet, the construction of the E-A model often assumes a uniform water level across the lake, thus overlooking potential segmentation during dry periods. To address this, our study implemented a zone-based approach, utilizing hydrological connectivity principles to ensure that elevation data within E-A models are confined to appropriate zonal regions. This method effectively minimized uncertainties by preventing errors from zonal discrepancies, significantly improving accuracy compared to traditional methods. It reduced root mean square errors (RMSE) by 0.71–1.73 m during the dry season, achieving RMSEs of 0.35, 0.64, and 0.37 m across three segments. Furthermore, this method ensures water level data are confined to specific zones, preventing the inconsistencies typically caused by averaging data across multiple stations or selecting data from varying elevations. This consistent domain definition reduces extrapolation errors during the model prediction and inversion. Moreover, the method compensates for time information losses often incurred by relying on multi-year percentile charts, thereby enabling more precise aquatic boundary delineation than traditional regional boundaries.

A predictive wear model in grinding using single-layer electroplated cBN tools

Wheel wear is a complex phenomenon significantly affecting grinding performance. While microscopic grit observation can detect various wear mechanisms, a predictive model is required to anticipate the total wear behavior under diverse grinding conditions. This study incorporates all wear mechanisms, including pullout, attrition, and fracture, into an analytical model aimed at determining improved grinding conditions. The grinding particles are approximated as standard constitutive geometries to model the wheel surface topography and the grit height reduction obtained from the wheel surface roughness is correlated with wear. The present model introduces an innovative approach to tracking and predicting wheel wear performance, eliminating the need for time-consuming grit micro-observation. Numerous grinding tests employing different tools were conducted, revealing strong agreement between the experimental results and the model predictions. The Application of this approach for predicting prolonged tool life is also demonstrated. Finally, the paper discusses how these findings contribute to a deeper understanding of the fundamental wear mechanisms in the grinding process.

Dispensability of extrinsic DnaA regulators in Escherichia coli cell-cycle control

Investigating a long-standing conceptual question in bacterial physiology, we examine why DnaA, the bacterial master replication initiator protein, exists in both ATP and ADP forms, despite only the ATP form being essential for initiation. We engineered the Δ4 Escherichia coli strain, devoid of all known external elements facilitating the DnaA-ATP/ADP conversion and found that these cells display nearly wild-type behaviors under nonoverlapping replication cycles. However, during rapid growth with overlapping cycles, Δ4 cells exhibit initiation instability. This aligns with our model predictions, suggesting that the intrinsic ATPase activity of DnaA alone is sufficient for robust initiation control in E. coli and the DnaA-ATP/ADP conversion regulatory elements extend the robustness to multifork replication, indicating an evolutionary adaptation. Moreover, our experiments revealed constant DnaA concentrations during steady-state cell elongation in both wild-type and Δ4 cells. These insights not only advance our understanding of bacterial cell-cycle regulation and DnaA but also highlight a fundamental divergence from eukaryotic cell-cycle controls, emphasizing protein copy-number sensing in bacteria versus programmed protein concentration oscillations in eukaryotes.

Coupled Multiphysics Modeling of Lithium-Ion Batteries for Automotive Crashworthiness Applications

Abstract Considerable advances have been made in battery safety models, but achieving predictive accuracy across a wide range of conditions continues to be challenging. Interactions between dynamically evolving mechanical, electrical, and thermal state variables make model prediction difficult during mechanical abuse scenarios. In this study, we develop a physics-based modeling approach that allows for choosing between different mechanical and electrochemical models depending on the required level of analysis. We demonstrate the use of this approach to connect cell-level abuse response to electrode-level and particle-level transport phenomena. A pseudo-two-dimensional model and simplified single-particle models are calibrated to electrical–thermal cycling data and applied to mechanically induced short-circuit scenarios to understand how the choice of electrochemical model affects the model prediction under abuse scenarios. These models are implemented using user-defined subroutines on ls-dyna finite element software and can be coupled with existing automotive crash safety models.

Capability and reliability of deep learning models to make density predictions on low-dose mammograms.

Breast density is associated with the risk of developing cancer and can be automatically estimated using deep learning models from digital mammograms. Our aim is to evaluate the capacity and reliability of such models to predict density from low-dose mammograms taken to enable risk estimates for younger women. We trained deep learning models on standard-dose and simulated low-dose mammograms. The models were then tested on a mammography dataset with paired standard- and low-dose images. The effect of different factors (including age, density, and dose ratio) on the differences between predictions on standard and low doses is analyzed. Methods to improve performance are assessed, and factors that reduce the model quality are demonstrated. We showed that, although many factors have no significant effect on the quality of low-dose density prediction, both density and breast area have an impact. The correlation between density predictions on low- and standard-dose images of breasts with the largest breast area is 0.985 (0.949 to 0.995), whereas that with the smallest is 0.882 (0.697 to 0.961). We also demonstrated that averaging across craniocaudal-mediolateral oblique (CC-MLO) images and across repeatedly trained models can improve predictive performance. Low-dose mammography can be used to produce density and risk estimates that are comparable to standard-dose images. Averaging across CC-MLO and model predictions should improve this performance. The model quality is reduced when making predictions on denser and smaller breasts.

Performance Prediction of GFRP-Reinforced Concrete Deep Beams Containing a Web Opening in the Shear Span

This study aimed to investigate the nonlinear structural behavior of concrete deep beams internally reinforced with glass fiber-reinforced polymer (GFRP) reinforcing bars and containing a web opening of various sizes and locations within the shear span. Three-dimensional (3D) numerical simulation models were developed for large-scale GFRP-reinforced concrete deep beams (300 mm × 1200 mm × 5000 mm) with a shear span-to-depth ratio (a/h) of 1.04. Predictions of the numerical models were validated against published experimental data. A parametric study was conducted to examine the effect of varying the opening size and location on the shear response. Results of the numerical analysis indicated that the strength of the deep beam models with an opening in the middle of the shear span decreased with an increase in either the opening width or height. The rate of the strength reduction caused by increasing the opening height was, however, more significant than that produced by increasing the opening width. Placing a web opening in the compression zone close to the load plate was very detrimental to the beam strength. Conversely, a negligible strength reduction was recorded when the web opening was placed in the tension side above the flexural reinforcement and away from the natural load path. Data of the parametric study were utilized to introduce simplified analytical formulas capable of predicting the shear capacity of GFRP-reinforced concrete deep beams with a web opening in the shear span.

Quasilinear theory and modelling of gyrokinetic turbulent transport in tokamaks

Abstract The theory, development, and validation of reduced quasilinear models of gyrokinetic turbulent transport in the closed flux surface core of tokamaks is reviewed. In combination with neoclassical collisional transport, these models are successful in accurately predicting core tokamak plasma temperature, density, rotation, and impurity profiles in a variety of confinement regimes. Refined experimental tests have been performed to validate the predictions of the quasilinear models, probing changes in the dominant gyrokinetic instabilities, as reflected in fluctuation measurements, cross-phases, and transport properties. These tests continue to produce a deeper understanding of the complex mix of instabilities at both electron and ion gyroradius scales.

Most general neutron decay correlations: Standard model recoil and radiative corrections

To continue from our previous work [], we derive the full Standard Model prediction of the most general free neutron differential decay rate with all massive particles (neutron, proton, and electron) polarized, including the O(1/mN) recoil corrections and O(α/π) radiative corrections. For the latter we adopt the newly developed pseudoneutrino formalism which is compatible to realistic experimental setups, in which neutrinos and photons are not detected. We also provide readily executable notebooks to evaluate these corrections. Published by the American Physical Society 2024

Advancing projections of crown-of-thorns starfish to support management interventions

Outbreaks of corallivorous Crown of Thorns Starfish (Acanthaster spp.; CoTS) cause substantial coral mortality throughout the Indo-Pacific, particularly on the Great Barrier Reef (GBR). Refining CoTS population density modelling and understanding the disparities between real-world observations and model predictions is crucial for developing effective control strategies. Using a spatially explicit ecosystem model of the GBR, we compared CoTS density model predictions to observations and incorporated a new zone-specific mortality rate to account for differences in predation of CoTS between fished and protected reefs. We found high congruence between predictions and observations: ∼81 % of categorical reef level CoTS densities matched or only differed by one category. However, underpredictions increased with higher observed densities. Zone-specific CoTS mortality reduced severe underpredictions from 7.1 % to 5.6 %, which is critical for managers as underpredictions indicate missing outbreaks where targeted culling is necessary, but also lead to underestimated coral loss attributed to CoTS outbreaks. Reef protection status affected prediction accuracy, highlighting the importance of further research on in situ CoTS mortality rates. The location of a reef inside or outside the “initiation box”, a speculative area of primary outbreaks (i.e., initial abrupt population increases) on the GBR, also influenced accuracy, with exact predictions more likely outside. Accurately modelling initiation box dynamics is challenging due to limited empirical data on CoTS outbreaks, highlighting the need for focussed research on outbreak dynamics to enhance predictive accuracy. Spatial factors, such as region and shelf position, contributed to the variance between observations and predictions, underscoring the importance of the spatial-temporal context of each observation. Observations of CoTS can help refine model predictions, guide targeted control measures, and contribute to effective ecosystem management for the long-term resilience of the GBR and other reefs targeted by CoTS throughout the Indo-Pacific.

Drug–target interaction prediction through fine-grained selection and bidirectional random walk methodology

The study of drug–target interaction plays an important role in the process of drug development. The subject of DTI forecasting has advanced significantly in the last several years, yielding numerous significant research findings and methodologies. Heterogeneous data sources provide richer information and comprehensive perspectives for drug–target interaction prediction, so many existing methods rely on heterogeneous networks, and graph embedding technology becomes an important technology to extract information from heterogeneous networks. These approaches, however, are less concerned with potential noisy information in heterogeneous networks and more focused on the extent of information extraction in those networks. Based on this, a potential DTI predictive network model called FBRWPC is proposed in this paper. It uses a fine-grained similarity selection program to first integrate similarity on similar networks and then a bidirectional random walk graph embedding learning method with restart to obtain an updated drug target interaction matrix. Through the use of similarity selection and fine-grained selection similarity integration, the framework can effectively filter out the noise present in heterogeneous networks and enhance the model's prediction performance. The experimental findings demonstrate that, even after being split up into four distinct types of data sets, FBRWPC can still retain great prediction performance, a sign of the model's resilience and good generalization.

Potential Controlling Factors and Landslide Susceptibility Features of the 2022 Ms 6.8 Luding Earthquake

On 5 September 2022, a Ms 6.8 earthquake struck Luding County, Ganzi Tibetan Autonomous Prefecture, Sichuan Province, China. This seismic event triggered over 16,000 landslides and caused serious casualties and infrastructure damages. The aim of this study is to perform the detailed landslides susceptibility mapping associated with this event based on an updated landslide inventory and logistic regression (LR) modeling. Firstly, we quantitatively assessed the importance of different controlling factors using the Jackknife and single-variable methods for modeling landslide occurrence. Subsequently, four landslide susceptibility assessment models were developed based on the LR model, and we evaluated the accuracy of the landslide susceptibility mappings using Receiver Operating Characteristic (ROC) curves and statistical measures. The results show that ground motion has the greatest influence on landslides in the entire study area, followed by elevation, while distance to rivers and topographic relief have little influence on the distribution of landslides. Compared to the NEE plate, PGA has a greater impact on landslides in the SWW plate. Moreover, the AUC value of the SWW plate significantly decreases for lithological types and aspect, indicating a more pronounced lithological control over landslides in the SWW plate. We attribute this phenomenon primarily to the occurrence of numerous landslides in Permian basalt and tuff in the SWW plate. Otherwise, the susceptibility results based on four models indicate that high-susceptibility areas predicted by different models are distributed along both sides of seismogenic faults and the Dadu Rivers. Landslide data have a significant impact on the model prediction results, and the model prediction accuracy based on the landslide data of the SWW plate is higher.

Generalized analytical description of relativistic strong-field ionization

A relativistic analytical theory of strong-field ionization applicable across the regimes of deep tunneling up to over-the-barrier ionization (OTBI) is developed, accounting also for the bound-state polarization and the Stark shift beyond perturbation theory. The latter improvement with respect to the state-of-the-art quasiclassical theory of Perelomov-Popov-Terent'ev (PPT) for strong-field ionization is essential to describe analytically the ionization in the OTBI regime and to resolve the order-of-magnitude discrepancy of the ionization yield in the relativistic regime with respect to PPT theory that has remained unexplained since the numerical result using the Klein-Gordon equation of Hafizi []. The predictions of the present relativistic model, in deviation to PPT theory, are shown to be observable using ultrashort laser pulses of relativistic intensities. Published by the American Physical Society 2024

Application of KNN algorithm incorporating Gaussian functions in green and high-quality development of cities empowered by circular economy

A growing number of industries have started to adapt to the circular economy since the concept's introduction. Therefore, in order to accurately evaluate the development level of circular economy, the circular economy prediction model based on support vector machine-Gaussian K-nearest neighbor is proposed. This model first uses the improved K-nearest neighbor algorithm based on Gaussian function to classify the index data of various levels, and then uses Support Vector Machine to make predictions based on relevant data. According to the experimental findings, the model's average prediction accuracy for each level of indicator was approximately 98.1%, 98.8%, 94.9%, and 95.9% for the levels of industrial development, resource consumption, ecological protection, and resource recycling and reuse, respectively. This prediction accuracy was higher than that of the multi-vector autoregressive model and the grey prediction model. The average prediction accuracy of the multi-vector autoregressive model, the grey prediction model, and the support vector machine-Gaussian K-nearest neighbor-based model in predicting the overall development level of the circular economy were about 94.3%, 96.2%, and 99.3%, respectively, with average recalls of about 86.6%, 87.7%, and 89.1%, and the average F1-measure was about 0.88, 0.89, and 0.92. Moreover, the average relative error based on the support vector machine-Gaussian K-nearest neighbour model was only approximately 0.6%, which was lower than the 3.7% and 2.8% for the multi-vector autoregressive model and the grey prediction model, respectively. Meanwhile, compared with the existing time series analysis techniques, the proposed SVM-Gaussian K nearest neighbor model fitted up to 0.95, which achieved good prediction performance. According to the above data, the support vector machine-Gaussian K-nearest neighbour model has the highest accuracy in predicting the amount of development of the circular economy.

A review of inverse problems for generalized elastic media: formulations, experiments, synthesis

AbstractStarting from the seminal works of Toupin, Mindlin and Germain, a wide class of generalized elastic models have been proposed via the principle of virtual work, by postulating expressions of the elastic energy enriched by additional kinematic descriptors or by higher gradients of the placement. More recently, such models have been adopted to describe phenomena which are not consistent with the Cauchy-Born continuum, namely the size dependence of apparent elastic moduli observed for micro and nano-objects, wave dispersion, optical modes and band gaps in the dynamics of heterogeneous media. For those structures the mechanical response is affected by surface effects which are predominant with respect to the bulk, and the scale of the external actions interferes with the characteristic size of the heterogeneities. Generalized continua are very often referred to as media with microstructure although a rigorous deduction is lacking between the specific microstructural features and the constitutive equations. While in the forward modelling predictions of the observations are provided, the actual observations at multiple scales can be used inversely to integrate some lack of information about the model. In this review paper, generalized continua are investigated from the standpoint of inverse problems, focusing onto three topics, tightly connected and located at the border between multiscale modelling and the experimental assessment, namely: (i) parameter identification of generalized elastic models, including asymptotic methods and homogenization strategies; (ii) design of non-conventional tests, possibly integrated with full field measurements and advanced modelling; (iii) the synthesis of meta-materials, namely the identification of the microstructures which fit a target behaviour at the macroscale. The scientific literature on generalized elastic media, with the focus on the higher gradient models, is fathomed in search of questions and methods which are typical of inverse problems theory and issues related to parameter estimation, providing hints and perspectives for future research.

COVID-19 pandemic-related drugs and microplastics from mask fibers jointly affect soil functions and processes.

The COVID-19 pandemic has led to an unprecedented increase in pharmaceutical drug consumption and plastic waste disposal from personal protective equipment. Most drugs consumed during the COVID-19 pandemic were used to treat other human and animal diseases. Hence, their nearly ubiquitous presence in the soil and the sharp increase in the last 3years led us to investigate their potential impact on the environment. Similarly, the compulsory use of face masks has led to an enormous amount of plastic waste. Our study aims to investigate the combined effects of COVID-19 drugs and microplastics from FFP2 face masks on important soil processes using soil microcosm experiments. We used three null models (additive, multiplicative, and dominative models) to indicate potential interactions among different pharmaceutical drugs and mask MP. We found that the multiple-factor treatments tend to affect soil respiration and FDA hydrolysis more strongly than the individual treatments. We also found that mask microplastics when combined with pharmaceuticals caused greater negative effects on soil. Additionally, null model predictions show that combinations of high concentrations of pharmaceuticals and mask MP have antagonistic interactions on soil enzyme activities, while the joint effects of low concentrations of pharmaceuticals (with or without MP) on soil enzyme activities are mostly explained by null model predictions. Our study underscores the need for more attention on the environmental side effects of pharmaceutical contamination and their potential interactions with other anthropogenic global change factors.

Modeling and mathematical analysis of root water uptake in the rhizosphere

ABSTRACT This research focuses on elucidating a system of partial differential equations designed to model the process of water uptake by a single root within an unsaturated soil. This model extends a unidimensional root scenario by incorporating radial and axial hydraulic conductance in the root structure. The flow dynamics in the rhizosphere are represented by the Richards equation, while a diffusion equation is employed to characterize water movement within the root. The resulting system forms a strongly coupled, nonlinear parabolic-elliptic system. To establish the existence and uniqueness of a solution, we utilize the technique of semi-discretization and the Schauder-Tychonoff fixed point theorem. Additionally, we approximate the system using a mixed finite element method, and the subsequent numerical tests presented align well with the model predictions. This comprehensive approach contributes to our understanding of the intricate dynamics involved in root water uptake and provides a robust framework for further exploration in this field.

Enhancing smart grid load forecasting: An attention-based deep learning model integrated with federated learning and XAI for security and interpretability

Smart grid is a transformative advancement that modernized the traditional power system for effective electricity management, and involves optimized energy distribution by load forecasting. Precise load forecasting provides the best utilization of energy resources and increases sustainability. Dynamic changes of several connected factors, such as temporal and geographical variability, pose challenges to accurate load prediction. Integrating Artificial Intelligence (AI) in the smart grid can enhance the performance of the forecasting process by capturing these changes. This study investigated load forecasting tasks on four different datasets. Several preprocessing and augmentation techniques are applied to increase the data quality. An attention-based 1D-CNN-GRU model is proposed to capture the temporal patterns from the time-series data, and the hyperparameters of the model are optimized using a particle swarm optimization (PSO) algorithm that also accelerates the convergence and results in an efficient training session. Empirical evaluations highlight that the proposed model substantially minimizes the loss, reflecting the ability to make accurate predictions. It obtains MAE values of 0.12, 0.8, 16.48, and 82.64 for the four datasets. Moreover, the explainable AI (XAI) technique is applied using Shapley Additive explanations (SHAP) to interpret the model prediction, providing the feature ranking based on their prediction score. Moreover, this study utilizes federated learning, enables collaborative training, maintains the privacy of the grid data, and secures the process comprehensively. The aggregation mechanism in federated learning is modified using pruning-based methods that reduce the parameters and computational cost, resulting in a more efficient framework. Integrating all these approaches provides valuable insights for developing a load forecasting model and outlines potential contributions in the smart grid domain.