Model Predictions Research Articles

Analysis of Molding Defection in IC Packaging and Testing Process

Molding injects a molding compound into a mold to form a protective shell around the wafer. During the injection process, overflow may occur, leading to mold flash, which reduces yield and causes significant manufacturing cost losses. This paper proposes a deep-learning-based method for detecting and predicting the occurrence of mold flash probability to address this issue. First, the paper conducts random forest importance analysis and correlation analysis to identify the key parameters that significantly impact mold flash. This paper uses these key parameters as input signals for the prediction model. The paper introduces an HLGA Transformer to construct an ensemble meta-learning model that predicts the probability of molding defects, achieving a prediction accuracy of 98.16%. The ensemble meta-learning approach proposed in this paper outperforms other methods in terms of performance. The model predictions can be communicated to the system in real time, allowing it to promptly adjust critical machine operation parameters, thereby significantly improving the molding process yield and reducing substantial manufacturing cost losses.

Causes of Multi-Mechanism Abnormal Formation Pressure in Offshore Oil and Gas Wells

This study thoroughly investigates the complex origins of abnormal formation pressure in offshore oil and gas wells, taking the Rio del Rey Basin in Cameroon as a case study. Renowned for its abundant oil and gas resources, the area faces unique challenges in predicting overpressure due to its high-temperature and high-pressure reservoir characteristics. By quantitatively analyzing the main mechanisms such as undercompaction, high-temperature fluid expansion, and mud diapirism, the study addresses the complexities of overpressure prediction. This paper introduces an innovative analytical framework that combines hierarchical clustering algorithms with the LightGBM model. Further refined by the application of Bayesian optimization, the model intelligently adjusts hyperparameters to enhance predictive accuracy. Utilizing well logging data and applying machine learning techniques, the paper classifies and identifies different mechanisms causing abnormal pressures, achieving a model prediction accuracy of 0.942. The research findings highlight the predominant role of the undercompaction mechanism, accounting for approximately 70% of the abnormal high-pressure events in the study area. Fluid expansion and shale diapirism contribute smaller but significant proportions of 10% and 20%, respectively. These quantitative insights into the pressure mechanisms are vital for optimizing drilling operations and reducing risks in oil and gas exploration. The study’s hybrid approach, integrating geophysical analysis with advanced computational techniques, sets a precedent for future research. It provides new avenues for applying machine learning to understand complex geological phenomena in similar geological environments and makes a significant contribution to the strategic planning of hydrocarbon exploration and production activities.

Non-destructive assessment of apple internal quality using rotational hyperspectral imaging

This work aims to predict the starch, vitamin C, soluble solids, and titratable acid contents of apple fruits using hyperspectral imaging combined with machine learning approaches. First, a hyperspectral camera by rotating samples was used to obtain hyperspectral images of the apple fruit surface in the spectral range of 380~1018 nm, and its region of interest (ROI) was extracted; then, the optimal preprocessing method was preferred through experimental comparisons; on this basis, genetic algorithms (GA), successive projection algorithms (SPA), and competitive adaptive reweighting adoption algorithms (CARS) were used to extract feature variables; subsequently, multiple machine learning models (support vector regression SVR, principal component regression PCR, partial least squares regression PLSR, and multiple linear regression MLR) were used to model the inversion between hyperspectral images and internal nutrient quality physicochemical indexes of fruits, respectively. Through the comparative analysis of all the model prediction results, it was found that among them, for starch, vitamin C, soluble solids, and titratable acid content, 2nd Der-CARS-MLR were the optimal prediction models with superior performance (the prediction coefficients of determination Rp2 exceeded 90% in all of them). In addition, potential relationships among four nutritional qualities were explored based on t-values and p-values, and a significant conclusion was drew that starch and vitamin C was highly correlated.

Chromosome-level reference genome for the Jonah crab, Cancer borealis.

The Jonah crab, Cancer borealis, is integral to marine ecosystems and supports a rapidly growing commercial fishery in the northwest Atlantic Ocean. This species also has a long history as a model for neuroscience that has expanded our understanding of central pattern generators, neuromodulation, synaptic plasticity, and the connectivity of neural circuits. Here we present a highly contiguous reference genome for the Jonah crab that will provide an essential resource to advance fisheries, conservation, and biomedical research. Using a combination of PacBio long-read sequencing and Omni-C scaffolding, we generated a final genome assembly spanning 691 Mb covering 51 chromosome-length scaffolds and 106 additional contigs. Benchmarking Universal Single-Copy Ortholog (BUSCO) analysis indicated a high-quality assembly with a completeness score of 90.8%. Repeat annotation identified 1,649 repeat families making up 48.27% of the Jonah crab genome. Gene model predictions annotated 24,830 protein coding genes with a 92.3% BUSCO score. Gene family evolution analysis revealed the expansion of gene families associated with nervous system function, and targeted analysis revealed an extensive repertoire of neural genes. The Jonah crab genome will not only provide a resource for neuroscience research but will also serve as a foundation to investigate adaptation to stress and population structure to support sustainable fisheries management during this time of rapidly changing environmental conditions in the northwest Atlantic Ocean.

Simple bounds on the most predictable component of a stochastic model

Special combinations of variables can have more predictability than any single variable in the combination. What is the maximum possible predictability that can be achieved through such combinations? Recently, this question was answered in the context of a linear stochastic model with fixed dynamics, where a sharp upper bound on predictability time was derived. However, the precise maximum is a complicated function of the entire spectrum of dynamical eigenvalues, obscuring any simple relation between predictability and eigenmodes. Based on numerical solutions of specific cases, it is conjectured here that the predictability of a stochastic model with a given least damped mode is bounded above by the predictability of a model in which all dynamical eigenvalues coalesce to the value corresponding to the least damped mode. Furthermore, it is shown that in this limit the maximum predictability time is determined by the largest root of a Laguerre polynomial. This result is used to prove the following simple bound: The maximum predictability time in the limit of coalesced eigenvalues is at most 4D-6 times the predictability time of the least damped mode, where D is the number of dynamical eigenmodes and D exceeds 3.

On the design of an optimal immigration policy

AbstractThis paper constructs a model for the study of optimal immigration from the perspective of natives. They have preferences over redistributive transfers, a public good subject to congestion, and over the level of family‐based migration. Border enforcement is costly. The model derives conditions for an optimal immigration policy, which balances skill‐based against family‐based migration, equalizes the fiscal opportunity cost of unauthorized immigration with the marginal cost of enforcement, and balances the effects on transfers against the implied congestion effects. A parameterized version of the model studies the 1994–2008 period and finds legislation projects on immigration consistent with the model's predictions.

Epidemiology of boxwood blight in western North Carolina and Virginia and evaluation of the boxwood blight infection risk model.

Boxwood blight, caused by Calonectria pseudonaviculata, is a highly invasive emerging disease. Since the first US report in North Carolina and Connecticut in 2011, boxwood blight has spread to over 30 US states, risking more than 90% of boxwood production. Our study investigated the disease field epidemiology and evaluated the boxwood blight infection risk model's prediction by analysing weekly blight monitoring data collected on detector plants exposed to the prevailing environmental conditions at two different locations (western Virginia and North Carolina) from spring through fall of 2014 to 2017. Boxwood blight was recorded in 61 of 86 weeks, with the highest infected leaf counts recorded in late summer or early fall. Rainfall, high relative humidity outside rainy periods and optimal temperatures (13.6-22.7°C) during prolonged leaf wetness (> 65h per week) had a significant positive effect on boxwood blight development. Classification analyses showed that disease predictions from the model using leaf wetness estimated by leaf wetness sensor were more closely aligned with observations from the field than predictions based on algorithms. This study improved our understanding of disease field epidemiology, provided leads to improve the existing model, and generated essential knowledge for formulating effective strategies for blight mitigation.

Knowledge-based inductive bias and domain adaptation for cell type annotation.

Measurement techniques often result in domain gaps among batches of cellular data from a specific modality. The effectiveness of cross-batch annotation methods is influenced by inductive bias, which refers to a set of assumptions that describe the behavior of model predictions. Different annotation methods possess distinct inductive biases, leading to varying degrees of generalizability and interpretability. Given that certain cell types exhibit unique functional patterns, we hypothesize that the inductive biases of cell annotation methods should align with these biological patterns to produce meaningful predictions. In this study, we propose KIDA, Knowledge-based Inductive bias and Domain Adaptation. The knowledge-based inductive bias constrains the prediction rules learned from the reference dataset, composed of multiple batches, to functional patterns relevant to biology, thereby enhancing the generalization of the model to unseen batches. Since the query dataset also contains gaps from multiple batches, KIDA's domain adaptation employs pseudo labels for self-knowledge distillation, effectively narrowing the distribution gap between model predictions and the query dataset. Benchmark experiments demonstrate that KIDA is capable of achieving accurate cross-batch cell type annotation.

Equivalence Ratio-Driven Flame Response of an Industrial Premixed Burner: Experiments and Modeling

Abstract This paper presents an analysis of the unsteady heat release rate response of premixed flames to equivalence ratio perturbations for an industrial premixed swirl-based burner. During this investigation, perfectly and technically premixed flames were acoustically forced via fuel/air mixture flow and air flow modulations, respectively, at the same operating conditions. From the resulting flame transfer functions (FTFs), measured using the multimicrophone method, the equivalence ratio driven FTF was isolated and extracted by removing the velocity driven component, i.e., the measured FTF from the perfectly premixed flame, from the technically premixed FTF with two novel extraction techniques. The results are compared with FTFs obtained directly in a previous experimental campaign where the fuel flow was acoustically forced, the resulting equivalence ratio fluctuations measured via an infrared absorption technique, and the heat release rate response to the forcing was quantified using chemiluminescence measurements. The results from both measurement approaches agreed well highlighting the validity of the techniques. Further, to understand the governing features of the equivalence ratio driven FTF, a physics-based analytical model following the G-equation approach was developed. The contributions from flame surface area, flame speed, and heat of reaction oscillations were modeled to describe the heat release rate dynamics. A limited number of physical parameters in the analytical model were anchored on one test condition, optimized and restricted to values, which were all physically reasonable, and were subsequently used for model predictions at other operating conditions. The FTF model predictions compared well with experimental data across a range of different operating conditions. Finally, the relative contributions from flame surface area, flame speed, and heat of reaction oscillations on the features of the FTFs were identified and explored.

Physics-Informed Data-Driven Approaches to State of Health Prediction of Maritime Battery Systems

Battery systems are increasingly being used for powering ocean going ships, and the number of fully electric or hybrid ships relying on battery power for propulsion and maneuvering is growing. In order to ensure the safety of such electric ships, it is of paramount importance to monitor the available energy that can be stored in the batteries, and classification societies typically require that the state of health (SOH) of the batteries can be verified by independent tests – annual capacity tests. However, this paper discusses physics-informed data-driven approaches to online diagnostics for state of health monitoring of maritime battery systems based on a combination of physical knowledge, physic-based models, insights from extensive characterization tests and operational sensor data collected from the batteries during actual operation. This represents an alternative approach to the annual capacity tests for electric ships that is found to be sufficiently robust and accurate under certain conditions. Previous attempts with purely data-driven models, including both cumulative and snapshot models, semi-supervised learning and simple models based on the state of charge did not achieve the required reliability and accuracy for them to be utilized in a ship classification perspective, as presented at previous PHM conferences. However, preliminary results from the physics-informed data-driven method presented in this paper indicate that it can be relied on for independent verification of state of health as an alternative to physical tests. It has been tested on battery cells cycled in laboratory degradation tests as well as on field returns from batteries onboard ships in service. Notwithstanding, further validation and verification of the method is recommended to further build confidence in the model predictions.

Tail probability bounds for acceptable random variables of partial sums and their application in linear models

Exponential probability inequalities play an important role in many parts of probability and statistics, offering a fundamental tool for tail probability bounding and modeling stochastic processes. The present work proposes a new exponential inequality for acceptable random variables. This class covers dependencies that are not embraced by the standard independence assumptions. Our results extend the theoretical perspective by placing sharper bounds on the above-mentioned variables, which tends to increase the accuracy of probabilistic modeling. Moreover, we study complete convergence in first-order autoregressive processes with acceptable innovations. The complete convergence framework gives further insights into the long-term behavior of the wide class of processes commonly used in time series analysis, econometrics, and other areas of applied research. In turn, our results on the influence of acceptable innovations on the properties of complete convergence establish bridges between theoretical probability inequalities and statistical applications. In addition, these contributions provide new tools in investigating the stability and predictability of the autoregressive models. These considerations consequently set up significant perspectives and enrich a developing body of research related to stochastic processes by reinforcing the theoretical framework as well as the methodological apparatus of probability theory.

Does the HCN/CO ratio trace the fraction of gravitationally bound gas? II. Variations in CO and HCN emissivity

We modeled emissivities of the HCN and CO $J=1-0$ transitions across a grid of molecular cloud models encapsulating observed properties that span from normal star-forming galaxies to more extreme merging systems. These models are compared with archival observations of the HCN and CO $J=1-0$ transitions, in addition to the radio continuum at 93 GHz, for ten nearby galaxies. We combined these model emissivities with the predictions of gravoturbulent models of star formation presented in the first paper in this series. In particular, we explored the impact of excitation and optical depth on CO and HCN emission and assess if the HCN/CO ratio tracks the fraction of gravitationally bound dense gas, $f_ grav $, in molecular clouds. We find that our modeled HCN/CO ratios are consistent with the measurements within our sample, and our modeled HCN and CO emissivities are consistent with the results of observational studies of nearby galaxies and clouds in the Milky Way. CO emission shows a wide range of optical depths across different environments, ranging from optically thick in normal galaxies to moderately optically thin in more extreme systems. HCN appears only moderately optically thick and shows significant subthermal excitation in both normal and extreme galaxies. We find an anticorrelation between HCN/CO and $f_ grav $, which implies that the HCN/CO ratio is not a reliable tracer of $f_ grav $. Instead, this ratio appears to best track gas at moderate densities ($n>10^ cm $), which is below the typically assumed dense gas threshold of $n>10^ cm $. We also find that variations in CO emissivity depend strongly on optical depth, which is a product of variations in the dynamics of the cloud gas. HCN emissivity is more strongly dependent on excitation, as opposed to optical depth, and thus does not necessarily track variations in CO emissivity. We further conclude that a single line ratio, such as HCN/CO, will not consistently track the fraction of gravitationally bound, star-forming gas if the critical density for star formation varies in molecular clouds. This work highlights important uncertainties that need to be considered when observationally applying an HCN conversion factor in order to estimate the dense (i.e.,\ cm $) gas content in nearby galaxies.

Non‐linear time history analyses of a rigid block isolated with unbonded fiber‐reinforced elastomeric isolators (UFREIs): A comparison between 3D finite element and phenomenological models

AbstractNumerical modeling represents a pivotal tool in the seismic analysis and design of structural systems, enabling the detailed prediction and examination of structural responses under seismic loading. This research conducts a comparative analysis of two numerical modeling approaches aimed at simulating the seismic response of unbonded fiber‐reinforced elastomeric isolators (UFREIs). The research focuses on a finite element (FE) model developed using Abaqus and a developed phenomenological model implemented in OpenSees, outlining the development and calibration processes for each. The FE model is developed based on simple rubber material testing data, while the phenomenological model is calibrated using experimental results from cyclic shear tests conducted on the UFREI device and the FE model. The primary objective of this study is to assess the effectiveness of these modeling approaches in predicting UFREI behavior under seismic conditions. This evaluation entails comparing model predictions with experimental data obtained from unidirectional shake table tests performed on a rigid block isolated by two UFREIs. This paper highlights the distinct advantages and limitations of each model in simulating UFREI dynamic responses during seismic events. Furthermore, it provides insights into the modeling techniques and discusses the computational demands and data requirements of each model, thereby aiding in their application to various aspects of seismic analysis and design.

Machine learning driven design of high-performance Al alloys

Aluminum (Al) alloys with both high strength and thermal conductivity (TC) are promising structural materials for wide application across different industries. Yet, design of such alloys is challenging, since strength and TC often share a trade-off. In this paper, we build prediction models for TC and ultimate tensile strength (UTS) of Al alloys using eXtreme gradient boosting (XGBoost) and support vector machine (SVM) algorithms, respectively. The models take physical descriptors from the alloy composition into account. Lasso and Gini Impurity algorithms were adopted for feature engineering. Guided by the models, an Al-2.64Si-0.43Mg-0.10Zn-0.03Cu alloy with TC over 190 W·m-1·K-1 and UTS over 220 MPa was designed. The alloy was fabricated and tested by experiment, and its UTS and TC are close to the model prediction. Microstructure characterization suggests that the fragmented and spherical Si phase, along with a few non-spherical Si phases, may be a key reason for the improved properties.

Explaining the US Health Paradox

Abstract Despite its high health spending, the US shows below-average life expectancy when compared to other OECD-countries. To find explanatory factors for this paradox, this article employs a Nested Analysis combining Large N Analysis (LNA) and Small N Analysis (SNA). First, it examines how far US life expectancy deviates from OLS model predictions which include a comprehensive set of health determinants. The results show that the established determinants clearly fail to explain the US underperformance. Second, a comparative analysis between the US and Canada identifies three idiosyncratic factors that explain the US deviation: its health care system particularities (1), the opioid crisis and drug-related deaths (2), and the particularly severe impact of the COVID-19 pandemic (3).

High frequency radar error classification and prediction based on K-means methods

This study aims to characterize the high frequency radar and numerically simulated low-frequency filtered currents in the south-eastern Bay of Biscay (study area) using a K-means classification algorithm based on an improved Euclidean Distance calculation method that does not take missing values. The errors between observations and simulations was estimated and predicted based on this classification method. Results indicate that predominantly eastward (northward) currents over the Spanish (French) continental shelf/slope in winter and more variable currents in the west and south-west in summer. The model classification results for circulation characteristics are in relatively good agreement with HF radar results, especially for currents on the Spanish (French) shelf/slope. In addition, the probabilistic relationship between observed and modeled currents was explored, obtaining the probability of occurrence of modeled current groups when each group of observed currents occurs. Finally, predictions of model and observed current errors were made based on the classification results, and it was found that the predictions based on the classification of all data had the smallest errors, with a 17% improvement over the unclassified control experiment. This study provides a foundation for subsequent model error testing, forecast product improvement and data assimilation.

Prediction and assessment of meteorological drought in southwest China using long short-term memory model

Abstract Drought prediction is crucial for mitigating risks and designing measures to alleviate its impact. Machine learning models have been widely applied in the field of drought prediction in recent years. This study concentrated on predicting meteorological droughts in southwest China, a region prone to frequent and severe droughts, particularly in areas with sparse meteorological station coverage. The long short-term memory (LSTM) predictive model, which is a deep learning model, was constructed by calculating standardized precipitation evapotranspiration index (SPEI) values based on 144 weather station observations from 1980 to 2020. The 5-fold cross-validation method was used for the hyperparameter optimization of the model. The LSTM model underwent comprehensive assessment and validation through multiple methods. This included the use of several accuracy assessment indicators and a comparison of results. The comparison covered different drought characteristics among the LSTM predictive model, the benchmark random forest (RF) predictive model, the historical drought situations, and the calculated SPEI values based on observations from 144 weather stations. The results showed that the training results of the LSTM predictive model basically agreed with the SPEI values calculated from weather station observations. The model-predicted variation trend of SPEI values for 2020 was similar to the variation in SPEI values calculated based on weather station observations. On the test set, the coefficient of determination (R 2), the root mean square error, the explained variance score, the Nash–Sutcliffe efficiency, and the Kling–Gupta efficiency were 0.757, 0.210, 0.802, 0.761, and 0.212, respectively. The total consistency rate of the drought grade was 59.26%. The spatial correlation distribution of SPEI values between LSTM model prediction and calculation from meteorological stations in 2020 was more than 0.5 for most regions. The correlation coefficients exceeded 0.6 in western Tibet and Chengdu Plains. Compared to the RF model, the LSTM model excelled in all five performance evaluation metrics and demonstrated a higher overall consistency rate for drought categories. The Kruskal–Wallis test for both the LSTM and RF models all indicated no significant difference in the distributions between the predicted and observed data. Scatter plots revealed that the prediction accuracy for both models in 2020 was suboptimal, with the SPEI showing a comparatively narrow range of values. Nonetheless, the LSTM model significantly outperformed the RF model in terms of prediction accuracy. In summary, the LSTM model demonstrated good overall performance, accuracy, and applicability. It has the potential to enhance dynamic drought prediction in regions with complex terrain, diverse climatic factors, and sparse weather station networks.

Enhanced Disease Detection for Apple Leaves with Rotating Feature Extraction

Leaf diseases such as Mosaic disease and Black Rot are among the most common diseases affecting apple leaves, significantly reducing apple yield and quality. Detecting leaf diseases is crucial for the prevention and control of these conditions. In this paper, we propose incorporating rotated bounding boxes into deep learning-based detection, introducing the ProbIoU loss function to better quantify the difference between model predictions and real results in practice. Specifically, we integrated the Plant Village dataset with an on-site dataset of apple leaves from an orchard in Weifang City, Shandong Province, China. Additionally, data augmentation techniques were employed to expand the dataset and address the class imbalance issue. We utilized the EfficientNetV2 architecture with inverted residual structures (FusedMBConv and S-MBConv modules) in the backbone network to build sparse features using a top–down approach, minimizing information loss. The inclusion of the SimAM attention mechanism effectively captures both channel and spatial attention, expanding the receptive field and enhancing feature extraction. Furthermore, we introduced depth-wise separable convolution and the CAFM in the neck network to improve feature fusion capabilities. Finally, experimental results demonstrate that our model outperforms other detection models, achieving 93.3% mAP@0.5, 88.7% Precision, and 89.6% Recall. This approach provides a highly effective solution for the early detection of apple leaf diseases, with the potential to significantly improve disease management in apple orchards.

Using CT images to assist the segmentation of MR images via generalization: Segmentation of the renal parenchyma of renal carcinoma patients.

Developing deep learning models for segmenting medical images in multiple modalities with less data and annotation is an attractive and challenging task, which was previously discussed as being accomplished by complex external frameworks for bridging the gap between different modalities. Exploring the generalization ability of networks in medical images in different modalities could provide more simple and accessible methods, yet comprehensive testing could still be needed. To explore the feasibility and robustness of using computed tomography (CT) images to assist the segmentation of magnetic resonance (MR) images via the generalization, in the segmentation of renal parenchyma of renal cell carcinoma (RCC) patients. Nephrographic CT images and fat-suppressed T2-weighted (fs-T2W) images were retrospectively collected. The pure CT dataset included 116 CT images. Additionally, 240 MR images were randomly divided into subsets A and B. From subset A, three training datasets were constructed, each containing 40, 80, and 120 images, respectively. Similarly, three datasets were constructed from subset B. Subsequently, datasets with mixed modality were created by combining these pure MR datasets with the 116 CT images. The 3D-UNET models for segmenting the renal parenchyma in two steps were trained using these 13 datasets: segmenting kidneys and then the renal parenchyma. These models were evaluated in internal MR (n=120), CT (n=65) validation datasets, and an external validation dataset of CT (n=79), using the mean of the dice similarity coefficient (DSC). To demonstrate the robustness of generalization ability over different proportions of modalities, we compared the models trained with mixed modality in three different proportions and pure MR, using repeated measures analysis of variance (RM-ANOVA). We developed a renal parenchyma volume quantification tool by the trained models. The mean differences and Pearson correlation coefficients between the model segmentation volume and the ground truth segmentation volume were calculated for its evaluation. The mean DSCs of models trained with 116 data in CT in the validation of MR were 0.826, 0.842, and 0.953, respectively, for the predictions of kidney segmentation model on whole image, renal parenchymal segmentation model on kidneys with RCC and without RCC. For all models trained with mixed modality, the means of DSC were above 0.9, in all validations of CT and MR. According to the results of the comparison between models trained with mixed modality and pure MR, the means of DSC of the former were significantly greater or equal to the latter, at all three different proportions of modalities. The differences of volumes were all significantly lower than one-third of the volumetric quantification error of a previous method, and the Pearson correlation coefficients of volumes were all above 0.96 on kidneys with and without RCC of three validations. CT images could be used to assist the segmentation of MR images via the generalization, with or without the supervision of MR data. This ability showed acceptable robustness. A tool for accurately measuring renal parenchymal volume on CT and MR images was established.

Optimal prediction and validation of Johnson-Cook failure model for Al 6061 using genetic algorithm

Optimal prediction and validation of Johnson-Cook failure model for Al 6061 using genetic algorithm