Superior Generalization Research Articles

Unveiling the Potential of Migrasomes: A Machine-Learning-Driven Signature for Diagnosing Acute Myocardial Infarction.

Recent studies have demonstrated that the migrasome, a newly functional extracellular vesicle, is potentially significant in the occurrence, progression, and diagnosis of cardiovascular diseases. Nonetheless, its diagnostic significance and biological mechanism in acute myocardial infarction (AMI) have yet to be fully explored. To remedy this gap, we employed an integrative machine learning (ML) framework composed of 113 ML combinations within five independent AMI cohorts to establish a predictive migrasome-related signature (MS). To further elucidate the biological mechanism underlying MS, we implemented single-cell RNA sequencing (scRNA-seq) of cardiac Cd45+ cells from AMI-induced mice. Ultimately, we conducted mendelian randomization (MR) and molecular docking to unveil the therapeutic effectiveness of MS. MS demonstrated robust predictive performance and superior generalization, driven by the optimal combination of Stepglm and Lasso, on the expression of nine migrasome genes (BMP1, ITGB1, NDST1, TSPAN1, TSPAN18, TSPAN2, TSPAN4, TSPAN7, TSPAN9, and WNT8A). Notably, ITGB1 was found to be predominantly expressed in cardiac macrophages in AMI-induced mice, mechanically regulating macrophage transformation between anti-inflammatory and pro-inflammatory. Furthermore, we showed a positive causality between genetic predisposition towards ITGB1 expression and AMI risk, positioning it as a causative gene. Finally, we showed that ginsenoside Rh1, which interacts closely with ITGB1, could represent a novel therapeutic approach for repressing ITGB1. Our MS has implications in forecasting and curving AMI to inform future diagnostic and therapeutic strategies for AMI.

Split-and-Fit: Learning B-Reps via Structure-Aware Voronoi Partitioning

We introduce a novel method for acquiring boundary representations (B-Reps) of 3D CAD models which involves a two-step process: it first applies a spatial partitioning , referred to as the "split", followed by a "fit" operation to derive a single primitive within each partition. Specifically, our partitioning aims to produce the classical Voronoi diagram of the set of ground-truth (GT) B-Rep primitives. In contrast to prior B-Rep constructions which were bottom-up, either via direct primitive fitting or point clustering, our Split-and-Fit approach is top-down and structure-aware , since a Voronoi partition explicitly reveals both the number of and the connections between the primitives. We design a neural network to predict the Voronoi diagram from an input point cloud or distance field via a binary classification. We show that our network, coined NVD-Net for neural Voronoi diagrams, can effectively learn Voronoi partitions for CAD models from training data and exhibits superior generalization capabilities. Extensive experiments and evaluation demonstrate that the resulting B-Reps, consisting of parametric surfaces, curves, and vertices, are more plausible than those obtained by existing alternatives, with significant improvements in reconstruction quality. Code will be released on https://github.com/yilinliu77/NVDNet.

Hybridized deep learning goniometry for improved precision in Ehlers-Danlos Syndrome (EDS) evaluation

BackgroundGeneralized Joint Hyper-mobility (GJH) can aid in the diagnosis of Ehlers-Danlos Syndrome (EDS), a complex genetic connective tissue disorder with clinical features that can mimic other disease processes. Our study focuses on developing a unique image-based goniometry system, the HybridPoseNet, which utilizes a hybrid deep learning model.ObjectiveThe proposed model is designed to provide the most accurate joint angle measurements in EDS appraisals. Using a hybrid of CNNs and HyperLSTMs in the pose estimation module of HybridPoseNet offers superior generalization and time consistency properties, setting it apart from existing complex libraries.MethodologyHybridPoseNet integrates the spatial pattern recognition prowess of MobileNet-V2 with the sequential data processing capability of HyperLSTM units. The system captures the dynamic nature of joint motion by creating a model that learns from individual frames and the sequence of movements. The CNN module of HybridPoseNet was trained on a large and diverse data set before the fine-tuning of video data involving 50 individuals visiting the EDS clinic, focusing on joints that can hyperextend. HyperLSTMs have been incorporated in video frames to avoid any time breakage in joint angle estimation in consecutive frames. The model performance was evaluated using Spearman’s coefficient correlation versus manual goniometry measurements, as well as by the human labeling of joint position, the second validation step.OutcomePreliminary findings demonstrate HybridPoseNet achieving a remarkable correlation with manual Goniometric measurements: thumb (rho = 0.847), elbows (rho = 0.822), knees (rho = 0.839), and fifth fingers (rho = 0.896), indicating that the newest model is considerably better. The model manifested a consistent performance in all joint assessments, hence not requiring selecting a variety of pose-measuring libraries for every joint. The presentation of HybridPoseNet contributes to achieving a combined and normalized approach to reviewing the mobility of joints, which has an overall enhancement of approximately 20% in accuracy compared to the regular pose estimation libraries. This innovation is very valuable to the field of medical diagnostics of connective tissue diseases and a vast improvement to its understanding.

Adaptive GN block-based model for seismic response prediction of train-bridge coupled systems

Deep learning (DL) methods has found extensive application in predicting structural responses. However, conventional neural networks often lack the capacity to effectively capture inter-data relationships, relying solely on training data for end-to-end prediction. Consequently, these methods may struggle with generalization, particularly when faced with new structural configurations absent from the training dataset. To overcome this limitation, this study adopts a graph-based framework, representing structures as graphs facilitates the transfer of features between different components, mimicking the transfer of forces in the real world. Subsequently, a Graph Neural Network Blocks (GN Blocks) model is introduced. Leveraging the holistic inference structure inherent in graph representations, the GN Block-based model as an adaptive model exhibits superior generalization performance in response analysis tasks involving unknown structural forms. Illustrated through the application to train-bridge coupled (TBC) systems, the GN Block-based model demonstrates high prediction accuracy and robust generalization performance, underscoring its suitability for tackling complex structural response prediction.

Detecting key genes relative expression orderings as biomarkers for machine learning-based intelligent screening and analysis of type 2 diabetes mellitus

Type 2 diabetes (T2D), a prevalent chronic metabolic disease of public health concern, is growing rapidly worldwide. In order to characterize robust transcriptomic differences between T2D and the healthy pancreatic islets, we proposed a novel data analysis approach based on the within-sample relative expression orderings of genes (REOs) and machine learning, which can efficiently integrate data from diverse gene expression datasets and help discover some potential signatures. We first identified the overlapped disease-specific reverse REOs from three bulk islets expression profiles that are substantially distinct. Subsequently, the minimum-Redundancy-Maximum-Relevance (mRMR) combined with incremental feature selection (IFS) strategy was applied to select the optimal reverse REOs by using the support vector machine (SVM). As a result, 7 optimal reverse REOs encompassing 12 genes were found, which could produce encouraging predictive performance on training data, independent testing, and external islets data, showcasing robust prediction capability and superior generalization despite the heterogeneity of data. Furthermore, the analysis of differential expression gene (DEG) in bulk islets data and pseudobulk method in single-cell RNA sequencing (scRNA-seq) islets data reaffirmed the crucial roles of these 7 REOs in islets function and disease progression. In conclusion, these REOs are robust biomarkers in T2D and may also be potential therapeutic targets for T2D. We hope these findings can help elucidate the underlying pathogenic mechanisms of T2D and contribute to the treatment of diabetes. The source code is available at https://github.com/xiexq007/T2D-REO.

Coarse to superfine: can hyperspectral soil organic carbon models predict higher-resolution information?

Introduction: Modeling and mapping of soil organic carbon concentration and distribution at the pedon scale is a current knowledge gap that can be addressed by laboratory-based hyperspectral imaging and chemometric analysis of soil cores. Despite the advancements in soil organic carbon models based on hyperspectral images, it is not clear how these models will perform upon input with images at higher resolutions than those of their training sets. This study aims to measure the generalizability of a soil organic carbon model based on a test set with higher resolution hyperspectral images than that of its training set.Methods: Organic carbon contents were measured at 10 cm intervals on eight soil cores for use as the training set and at 1 cm intervals on a single core for use as the test set. Three regression models, namely, multilayer perceptron, partial least-squares, and support vector regressions, were trained and tested with the median of each hyperspectral image for each of these intervals as the training and test predictors. Permutation importance analysis was performed to explain the models.Results: The results show that although all three models had the same validation R2 of 0.92 for cross-validation on the 10 cm data, multilayer perceptron regression allowed the best generalization with a test R2 of 0.96 compared to the partial least-squares regression (0.81) and support vector regression (0.86). It was demonstrated that the multilayer perceptron model is more robust to soil surface anomalies and that it predicts soil organic carbon on the test set by learning the spectral features related to soil organic matter chromophore activity in the 950–1,150 nm region along with clay mineralogy derived from peaks at 1,400, 1,900, 2,200, 2,250, and 2,350 nm.Conclusions: This study shows that while the regression models based on hyperspectral images perform well at the 10-cm-resolution cross validation, multilayer perceptron regression shows superior generalization and robustness for a higher 1-cm-resolution test set without much loss of prediction power.

DdeNet: A dual-domain end-to-end network combining Pale-Transformer and Laplacian convolution for sparse view CT reconstruction

Sparse view (SV) computed tomography (CT) is a clinical diagnostic technique aimed at reducing radiation dose to the human body from X-rays. However, SVCT reconstructed using conventional filtered back-projection (FBP) algorithms will produce severe streak artifacts. In recent years, convolutional neural networks (CNNs) have demonstrated considerable success in addressing the inverse problem of SVCT reconstruction. However, the local convolutional operations of CNNs ignores long-range dependencies of contextual feature information, thereby limiting their capability in feature extraction. To address this limitation, we propose a dual-domain end-to-end network (DdeNet) combining Pale-Transformer and Laplacian convolution for SVCT reconstruction. Initially, in the projection domain, a UNet is employed to restore the interpolated sinogram. Subsequently, in the image domain, we introduce Pale-Transformer and Laplacian convolution to design a dual-stream feature fusion network for finely restoring the CT images reconstructed by FBP. The dual-stream feature fusion network comprises an image restoration network (IRN) module and an edge enhancement network (EEN) module. The depthwise separable convolution based modified pale-shaped self-attention mechanism in Pale-Transformer captures global feature information of the CT image in the IRN, while the Laplacian convolution and multi-scale convolution block with a squeeze and excitation mechanism in the EEN further extract and restore edge feature information of the CT image. Finally, the feature information output from IRN and EEN is fused along the channel dimension and then convolved to obtain the final reconstructed CT image. Experimental results demonstrate that DdeNet exhibits superior generalization and reconstruction performance compared to the previous CNN or Transformer-based reconstruction methods.

FDiff-Fusion: Denoising diffusion fusion network based on fuzzy learning for 3D medical image segmentation

In recent years, the denoising diffusion model has achieved remarkable success in image segmentation modeling. With its powerful nonlinear modeling capabilities and superior generalization performance, denoising diffusion models have gradually been applied to medical image segmentation tasks, bringing new perspectives and methods to this field. However, existing methods overlook the uncertainty of segmentation boundaries and the fuzziness of regions, resulting in the instability and inaccuracy of the segmentation results. To solve this problem, a denoising diffusion fusion network based on fuzzy learning for 3D medical image segmentation (FDiff-Fusion) is proposed in this paper. By integrating the denoising diffusion model into the classical U-Net network, this model can effectively extract rich semantic information from input medical images, thus providing excellent pixel-level representation for medical image segmentation. In this paper, a fuzzy learning module is designed on the skip path of U-Net network because of the widespread boundary uncertainty and region blurring of medical image segmentation. The module sets several fuzzy membership functions for the input encoded features to describe the similarity degree between the feature points, and applies fuzzy rules to the fuzzy membership functions, thus enhancing the modeling ability of the model for uncertain boundaries and fuzzy regions. In addition, in order to improve the accuracy and robustness of the model segmentation results, we introduced an iterative attention feature fusion method in the test phase, which added local context information to the global context information in the attention module to fuse the prediction results of each denoising time step. Finally, to validate the effectiveness of FDiff-Fusion, we compare it with existing advanced segmentation networks on the BRATS 2020 brain tumor dataset and the BTCV abdominal multi-organ dataset. The results show that FDiff-Fusion significantly improves the Dice scores and HD95 distance on these two datasets, demonstrating its superiority in medical image segmentation tasks.

Intuitionistic fuzzy generalized eigenvalue proximal support vector machine

Generalized eigenvalue proximal support vector machine (GEPSVM) has attracted widespread attention due to its simple architecture, rapid execution, and commendable performance. GEPSVM gives equal significance to all samples, thereby diminishing its robustness and efficacy when confronted with real-world datasets containing noise and outliers. In order to reduce the impact of noises and outliers, we propose a novel intuitionistic fuzzy generalized eigenvalue proximal support vector machine (IF-GEPSVM). The proposed IF-GEPSVM assigns the intuitionistic fuzzy score to each training sample based on its location and surroundings in the high-dimensional feature space by using a kernel function. The solution of the IF-GEPSVM optimization problem is obtained by solving a generalized eigenvalue problem. Further, we propose an intuitionistic fuzzy improved generalized eigenvalue proximal support vector machine (IF-IGEPSVM) by solving standard eigenvalue decomposition resulting in simpler optimization problems with less computation cost which leads to an efficient intuitionistic fuzzy-based model. We conduct a comprehensive evaluation of the proposed IF-GEPSVM and IF-IGEPSVM models on UCI and KEEL benchmark datasets. Moreover, to evaluate the robustness of the proposed IF-GEPSVM and IF-IGEPSVM models, label noise is introduced into some UCI and KEEL datasets. The experimental findings showcase the superior generalization performance of the proposed IF-GEPSVM and IF-IGEPSVM models when compared to the existing baseline models, both with and without label noise. Our experimental results, supported by rigorous statistical analyses, confirm the superior generalization abilities of the proposed IF-GEPSVM and IF-IGEPSVM models over the baseline models. Furthermore, we implement the proposed IF-GEPSVM and IF-IGEPSVM models on the USPS recognition dataset, yielding promising results that underscore the models’ effectiveness in practical and real-world applications. The source code of the proposed IF-GEPSVM and IF-IGEPSVM models are available at https://github.com/mtanveer1/IF-GEPSVM.

Detection Based on Semantics and a Detail Infusion Feature Pyramid Network and a Coordinate Adaptive Spatial Feature Fusion Mechanism Remote Sensing Small Object Detector

In response to the challenges of remote sensing imagery, such as unmanned aerial vehicle (UAV) aerial imagery, including differences in target dimensions, the dominance of small targets, and dense clutter and occlusion in complex environments, this paper optimizes the YOLOv8n model and proposes an innovative small-object-detection model called DDSC-YOLO. First, a DualC2f structure is introduced to improve the feature-extraction capabilities of the model. This structure uses dual-convolutions and group convolution techniques to effectively address the issues of cross-channel communication and preserving information in the original input feature mappings. Next, a new attention mechanism, DCNv3LKA, was developed. This mechanism uses adaptive and fine-grained information-extraction methods to simulate receptive fields similar to self-attention, allowing adaptation to a wide range of target size variations. To address the problem of false and missed detection of small targets in aerial photography, we designed a Semantics and Detail Infusion Feature Pyramid Network (SDI-FPN) and added a dedicated detection scale specifically for small targets, effectively mitigating the loss of contextual information in the model. In addition, the coordinate adaptive spatial feature fusion (CASFF) mechanism is used to optimize the original detection head, effectively overcoming multi-scale information conflicts while significantly improving small target localization accuracy and long-range dependency perception. Testing on the VisDrone2019 dataset shows that the DDSC-YOLO model improves the mAP0.5 by 9.3% over YOLOv8n, and its performance on the SSDD and RSOD datasets also confirms its superior generalization capabilities. These results confirm the effectiveness and significant progress of our novel approach to small target detection.

Fire detection and risk assessment via Support Vector Regression with Flattening-Samples Based Augmented Regularization

We propose a Hybrid Support Vector Regression (SVR) with Flattening-Samples Based Augmented Regularization (Hybrid FSR-SVR) architecture for multi-sensor fire detection and forest fire risk assessment. The Hybrid FSR-SVR is a lightweight architecture built upon the novel Flattening-Samples Based Augmented Regularization (FSR) approach and temporal trends of environmental variables. The FSR approach augments l2 norm based smoothing term into an l1-l2 combination, facilitating the integration of l1 regularization into the SVR method, thereby enhancing generalization with minimal computational load. We evaluate the performance of Hybrid FSR-SVR using two distinct datasets covering indoor and forest fires, benchmarking against 15 machine learning models, including state-of-the-art techniques, such as Recurrent Trend Predictive Neural Network (rTPNN), Long-Short Term Memory (LSTM), Multi-Layer Perceptron (MLP), Gated Recurrent Unit (GRU), and Gradient Boosting. Our findings demonstrate that Hybrid FSR-SVR effectively assesses the risk of forest fire, enabling early preventive measures. Notably, it achieves a remarkable accuracy of 0.95 for forest fire detection and ranks third with 0.88 accuracy for indoor fire detection. Importantly, it exhibits computation times significantly lower – by 1 to 2 orders of magnitude – than the majority of compared techniques. The superior generalization ability of Hybrid FSR-SVR, facilitated by flattening-samples based augmented regularization, allows for high detection performance even with smaller training sets.

An efficient transformer network for detecting multi-scale chicken in complex free-range farming environments via improved RT-DETR

Accurate detection and localization of chicken targets in complex environments are the foundation of intelligent free-range chicken farming. However, commercial free-range chicken farming poses several challenges to detection methods, including complex and variable backgrounds, multi-scale targets, aggregated occlusion and changes in the posture of chicken. While there has been significant research on object detection algorithms based on convolutional neural networks, these algorithms have limited effectiveness in detecting small targets and generalizing to new scenes. Therefore, this study proposes a highly accurate and strongly generalizable chicken object detection model called Efficient Multi-scale Chicken DETR (EMSC-DETR), which is based on an improved Real-Time DEtection TRansformer (RT-DETR). To address the issue of easily losing small target features, a new efficient transformer module called space-to-depth transformer module (SDTM) is introduced. SDTM includes space-to-depth (SPD) and BiFormer modules, which facilitate deep interaction between local and global features, and significantly improve the computational efficiency of the transformer. Additionally, to overcome the detection and differentiation difficulties caused by chicken aggregation occlusion, the transformer encoder of EMSC-DETR incorporates a contextual transformer (CoT). CoT utilizes rich contextual information among the nearest neighbors to enhance the feature representation of high-level features, enabling more accurate localization of overlapped chicken targets. To evaluate the proposed model in complex environments, 14 scene data were collected from two fields and divided into training, validation, testing sets and three independent testing sets with unknown scenes. Experimental results show that the proposed EMSC-DETR obtains the best detection performance in the testing set, with an AP50 of 98.6%, surpassing other state-of-the-art models of similar size. In the three independent testing sets, the proposed model shows improvements of 8.2%, 9.2% and 28.3% in the AP50 compared to the original RT-DETR, respectively. Moreover, the model exhibits superior generalization ability for unknown scenes compared to other state-of-the-art models. Robustness and fine-tuning experiments further confirm the reliability and application potential of EMSC-DETR in detecting free-range chickens, which is crucial for complex commercial farming. This study is expected to provide references and new ideas for precision livestock farming.

Enhancing Medical Imaging Segmentation with GB-SAM: A Novel Approach to Tissue Segmentation Using Granular Box Prompts.

Recent advances in foundation models have revolutionized model development in digital pathology, reducing dependence on extensive manual annotations required by traditional methods. The ability of foundation models to generalize well with few-shot learning addresses critical barriers in adapting models to diverse medical imaging tasks. This work presents the Granular Box Prompt Segment Anything Model (GB-SAM), an improved version of the Segment Anything Model (SAM) fine-tuned using granular box prompts with limited training data. The GB-SAM aims to reduce the dependency on expert pathologist annotators by enhancing the efficiency of the automated annotation process. Granular box prompts are small box regions derived from ground truth masks, conceived to replace the conventional approach of using a single large box covering the entire H&E-stained image patch. This method allows a localized and detailed analysis of gland morphology, enhancing the segmentation accuracy of individual glands and reducing the ambiguity that larger boxes might introduce in morphologically complex regions. We compared the performance of our GB-SAM model against U-Net trained on different sizes of the CRAG dataset. We evaluated the models across histopathological datasets, including CRAG, GlaS, and Camelyon16. GB-SAM consistently outperformed U-Net, with reduced training data, showing less segmentation performance degradation. Specifically, on the CRAG dataset, GB-SAM achieved a Dice coefficient of 0.885 compared to U-Net's 0.857 when trained on 25% of the data. Additionally, GB-SAM demonstrated segmentation stability on the CRAG testing dataset and superior generalization across unseen datasets, including challenging lymph node segmentation in Camelyon16, which achieved a Dice coefficient of 0.740 versus U-Net's 0.491. Furthermore, compared to SAM-Path and Med-SAM, GB-SAM showed competitive performance. GB-SAM achieved a Dice score of 0.900 on the CRAG dataset, while SAM-Path achieved 0.884. On the GlaS dataset, Med-SAM reported a Dice score of 0.956, whereas GB-SAM achieved 0.885 with significantly less training data. These results highlight GB-SAM's advanced segmentation capabilities and reduced dependency on large datasets, indicating its potential for practical deployment in digital pathology, particularly in settings with limited annotated datasets.

A multi-stage spatial queueing model with logistic arrivals and departures consistent with the microscopic fundamental diagram and hysteresis

This paper introduces a spatial queueing model for a single bottleneck during morning peak hours. Utilizing the logistic function and after appropriate calibration, it articulates the arrival and departure flows in continuous, differentiable terms. By validating the model across different peak periods and locations, the demand model’s robustness is superior to other commonly used functions. This model also incorporates constant or varying capacity scenarios. It effectively captures key aspects of morning peak traffic, including the emergence of hysteresis loops in fundamental diagrams (FDs) of density and flow. The model’s multi-stage approach recognizes three distinct phases in traffic flow: freeflow, transition, and queued segments, ensuring spatial consistency in flow and density across these stages. It accounts for the growth of the queued segment and vehicle spillback under various bottleneck intensities, with the resulting FDs for speed and density also displaying hysteresis loops. The calibration of model parameters utilizes time-series data of traffic flow and density space–time maps derived from real-world data. The validation results accurately reflect real traffic scenarios, emulating the counterclockwise hysteresis loops observed in density and its heterogeneity, and provide both planar and three-dimensional FDs at different points along the traffic link, each mirroring real-life traffic patterns. Additionally, a comparison with the cell transmission model (CTM) reveals that the proposed model exhibits superior generalization and robustness.

GROM: A generalized routing optimization method with graph neural network and deep reinforcement learning

Routing optimization, as a significant part of Traffic Engineering (TE), plays an important role in balancing network traffic and improving quality of service. With the application of Machine Learning (ML) in various fields, many neural network-based routing optimization solutions have been proposed. However, most existing ML-based methods need to retrain the model when confronted with a network unseen during training, which incurs significant time overhead and response delay. To improve the generalization ability of the routing model, in this paper, we innovatively propose a routing optimization method GROM which combines Deep Reinforcement Learning (DRL) and Graph Neural Networks (GNN), to directly generate routing policies under different and unseen network topologies without retraining. Specifically, for handling different network topologies, we transform the traffic-splitting ratio into element-level output of GNN model. To make the DRL agent easier to converge and well generalize to unseen topologies, we discretize the huge continuous traffic-splitting action space. Extensive simulation results on five real-world network topologies demonstrate that GROM can rapidly generate routing policies under different network topologies and has superior generalization ability.

Probabilistic assessment of rockburst risk in TBM-excavated tunnels with multi-source data fusion

In recent years, tunnel boring machines (TBMs) have become extensively utilized in underground engineering projects. However, as a prevalent type of dynamic geological hazard, rockburst poses a serious threat to the safety of personnel, equipment, and property. To ensure the safe advancement of TBMs, we integrate the data of microseismic monitoring, TBM excavation, and surrounding rock and establish multiple probabilistic assessment models of rockburst risk using probabilistic ensemble learning and quantum particle swarm optimization. In order to identify the most optimal model for geological engineers, a comprehensive evaluation system is devised based on metrics like accuracy, Cohen’s kappa, and F1-score. This system provides a thorough assessment of the models’ global and local generalization performance. The results indicate that the classification and regression tree-extremely randomized trees (CART-ERT) hybrid model achieves the highest score, demonstrating superior generalization performance with an accuracy of 93.06% and a Cohen’s kappa of 0.9074. Moreover, the F1-scores for no rockburst, low rockburst, moderate rockburst, and strong rockburst are 0.9412, 0.9444, 0.9189, and 0.9189, respectively. Based on the operational framework of the CART-ERT hybrid model, a user-friendly graphical user interface (GUI) system is developed. This significantly enhances the practicality and deployability of the model. Through application in a TBM diversion tunnel project in Xinjiang, the on-site early-warning accuracy of rockburst with the GUI system reaches 90%. Notably, the GUI system maintains high early-warning sensitivity and reliability for high-intensity rockburst such as moderate and strong ones. Lastly, to enhance the model’s interpretability, a variable importance measure (VIM) analysis on 14 features extracted from three types of heterogeneous data is conducted to assess their contributions to the early warning of rockburst. We discover that the cumulative energy change rate of microseismic events exhibits the highest importance score, indicating its crucial role in rockburst early warning.

Developing Models Inspired by CNN and DNN for more Effective Precipitation Forecasting Techniques

The study aims to analyze some changes in rainfall behavior in (10) global cities, including Tokyo, Osaka, Shanghai, Chicago, Istanbul, Buenos Aires, Amsterdam, Madrid, Florance, and New York using an advanced machine learning (ML) approach for predicting the average precipitation. Utilizing the Deep Neural Network (DNN) and Convolution Neural Network (CNN) model, we have acquired mean precipitation monthly data from the selected cities with covering the period of 2000-2023. In order to assess mean precipitation in cities, we effectively employed both the average mutual information (AMI) during the model’s development. Constructed using both a training subset (80% of data from 2000 to 2018) and a testing subset (20% of data from 2019 to 2023), MP forecasting models for ten cities were established. When compared to the DNN model, the CNN model truly shined by showcasing its remarkable potential as a high-level model in accurately estimating mean precipitation values of cities while demonstrating superior generalization ability and minimal variance.

Predictive Study on the Occurrence of Wheat Blossom Midges Based on Gene Expression Programming with Support Vector Machines.

This study addresses the challenges in plant pest and disease prediction within the context of smart agriculture, highlighting the need for efficient data processing techniques. In response to the limitations of existing models, which are characterized by slow training speeds and a low prediction accuracy, we introduce an innovative prediction method that integrates gene expression programming (GEP) with support vector machines (SVM). Our approach, the gene expression programming-support vector machine (GEP-SVM) model, begins with encoding and fitness function determination, progressing through cycles of selection, crossover, mutation, and the application of a convergence criterion. This method uniquely employs individual gene values as parameters for SVM, optimizing them through a grid search technique to refine genetic parameters. We tested this model using historical data on wheat blossom midges in Shaanxi Province, spanning from 1933 to 2010, and compared its performance against traditional methods, such as GEP, SVM, naive Bayes, K-nearest neighbor, and BP neural networks. Our findings reveal that the GEP-SVM model achieves a leading back-generation accuracy rate of 90.83%, demonstrating superior generalization and fitting capabilities. These results not only enhance the computational efficiency of pest and disease prediction in agriculture but also provide a scientific foundation for future predictive endeavors, contributing significantly to the optimization of agricultural production strategies.

Moving toward Automaticity: A Robust Synthetic Occlusion Image Method for High-Throughput Mushroom Cap Phenotype Extraction

Mushroom cap is a key trait in the growth process and its phenotypic parameters are essential for automatic cultivation and smart breeding. However, the edible mushrooms are usually grown densely with mutual occlusion, which is difficult to obtain the phenotypic parameters non-destructively. Although deep learning methods achieve impressive performance with superior generalization capabilities, they require a large amount of ground truth label of the occluded target, which is a challenging task due to the substantial workload. To overcome this obstacle, a novel synthetic cap occlusion image method was proposed for rapidly generating edible mushroom occlusion datasets using raw images, in which the ground truth is obtained from the real world and the occlusion is randomly generated for simulating real scenes. Moreover, variants of amodal instance segmentation models with different backbone were trained and evaluated on our synthetic occlusion image datasets. Finally, an amodal mask-based size estimation method was presented to calculate the width and length of the cap. The experimental results showed that the amodal instance segmentation achieved an AP@[0.5:0.95] of 82%, 93% and 96% on Oudemansiella raphanipes, Agrocybe cylindraceas and Pholiota nameko synthetic cap datasets, respectively, with a size of 1024 × 1024 px, which indicates that our occlusion image synthesis method can effectively simulate the real cap occlusion situation. The size estimation method achieved an R2 of 0.95 and 0.98 between predictive amodal caps and manually labeled caps for the length and width of Agrocybe cylindraceas cap, respectively, which can be applied to obtain the phenotypic parameters of each cap effectively and accurately. These methods not only meet the demand for automatic monitoring of edible mushroom morphology in factories but also provide technical support for intelligent breeding.

Class-agnostic counting and localization with feature augmentation and scale-adaptive aggregation

Compared to traditional class-specific counting, class-agnostic counting (CAC) enables the counting of arbitrary novel categories with only a few exemplars. The core issue of CAC lies in feature matching between the query image and exemplars. However, the scarce user-given exemplars make it difficult to adequately cover the diverse appearance of target objects, which poses challenges for feature matching. In this paper, we propose a Class-Agnostic Counting and Localization model (CACL) to enrich the exemplars by mining more inherent instances from the query feature. Specifically, CACL consists mainly of Self-Augment Feature Matching (SAFM) and Scale-Adaptive Spatial Aggregation (SASA). SAFM collects features of target instances frequently occurring in the query feature to augment the limited user-given exemplar, thus producing more comprehensive semantic features of the target category. To improve the robustness against scale variation, we design the SASA to enable the model to aggregate multiscale features adaptively. Compared to previous density map-based approaches, our method can locate the target objects which is more practical in real-world applications. Extensive experiments on FSC-147 suggest that our method achieves a competitive counting performance compared to state-of-the-art methods. Moreover, cross-dataset experiments demonstrate superior generalization in class-specific counting tasks.