Runtime Efficiency Research Articles

An advanced framework for net electricity consumption prediction: Incorporating novel machine learning models and optimization algorithms

In recent years, the escalating demand for electric energy has underscored the need for robust prediction models capable of accurately anticipating consumption patterns. The imperative lies in enabling utilities and policymakers to optimize resource allocation, strategically plan infrastructure development, and ensure the stability and efficiency of the power grid. This study undertakes a comprehensive comparative analysis of machine learning techniques employed in predicting net electricity consumption in Turkey. The primary goal is to augment the accuracy and performance of electricity load forecasting, thereby contributing to effective energy management and fostering sustainable development within the power sector. Two machine learning models, including CatBoost and Extreme Gradient Boosting (XGBoost), are strategically integrated with optimization algorithms such as Sparrow Search Algorithm (SSA), Phasor Particle Swarm Optimization (PPSO), and Hybrid Grey Wolf Optimization (GWO). The core analysis centers on evaluating the performance of these integrated models based on key accuracy metrics and runtime efficiency. Notably, the results underscore that the XGBoost-SSA model emerges as the superior performer, exhibiting heightened accuracy and superior performance in predicting electricity consumption. This model showcases the highest coefficient of determination (R2) value and demonstrates lower errors during the testing phase, thereby presenting a promising and effective approach for electricity consumption prediction in the specific context of Turkey.

Practical multi-party private set intersection cardinality and intersection-sum protocols under arbitrary collusion1

Private set intersection cardinality (PSI-CA) and private intersection-sum with cardinality (PSI-CA-sum) are two primitives that enable data owners to learn the intersection cardinality of their data sets, with the difference that PSI-CA-sum additionally outputs the sum of the associated integer values of all the data that belongs to the intersection (i.e., intersection-sum). However, to the best of our knowledge, all existing multi-party PSI-CA (MPSI-CA) protocols are either limited by high computational cost or face security challenges under arbitrary collusion. As for multi-party PSI-CA-sum (MPSI-CA-sum), there is even no formalization for this notion at present, not to mention secure constructions for it. In this paper, we first present an efficient MPSI-CA protocol with two non-colluding parties. This protocol significantly decreases the number of parties involved in expensive interactive procedures, leading to a significant enhancement in runtime efficiency. Our numeric results demonstrate that the running time of this protocol is merely one-quarter of the time required by our proposed MPSI-CA protocol that is secure against arbitrary collusion. Therefore, in scenarios where performance is a priority, this protocol stands out as an excellent choice. Second, we successfully construct the first MPSI-CA protocol that achieves simultaneous practicality and security against arbitrary collusion. Additionally, we also conduct implementation to verify its practicality (while the previous results under arbitrary collusion only present theoretical analysis of performance, lacking real implementation). Numeric results show that by shifting the costly operations to an offline phase, the online computation can be completed in just 12.805 seconds, even in the dishonest majority setting, where 15 parties each hold a set of size 2 16 . Third, we formalize the concept of MPSI-CA-sum and present the first realization that ensures simultaneous practicality and security against arbitrary collusion. The computational complexity of this protocol is roughly twice that of our MPSI-CA protocol. Besides the main results, we introduce the concepts and efficient constructions of two novel building blocks: multi-party secret-shared shuffle and multi-party oblivious zero-sum check, which may be of independent interest.

Misbehavior detection with spatio-temporal graph neural networks

Graph Neural Networks (GNNs) gained the attention of researchers following advancements in Representational Learning. Unlike classical machine learning (ML) methods, its ability to represent more concepts on its data structure gives GNN a clear head start. On the other hand, Misbehavior Detection (MBD) has become a security solution for authorized vehicles to protect against malicious activities on Vehicles to Everything (V2X) Networks. Although authorities standardize MBD, there are not enough MBD datasets for ML, and existing ones are unsuitable for GNNs. In this study, we address this issue by providing an algorithm to convert classical MBD datasets for GNN. Then, a novel GNN model is proposed to detect misbehaving activity on V2X networks. Obtained results show that the proposed GNN model outperforms existing methods with 99.92% accuracy, 99.9196% recall, 31% better runtime efficiency, and 97.35% Matthews correlation coefficient.

Affine medical image registration with fusion feature mapping in local and global.

Objective. Medical image affine registration is a crucial basis before using deformable registration. On the one hand, the traditional affine registration methods based on step-by-step optimization are very time-consuming, so these methods are not compatible with most real-time medical applications. On the other hand, convolutional neural networks are limited in modeling long-range spatial relationships of the features due to inductive biases, such as weight sharing and locality. This is not conducive to affine registration tasks. Therefore, the evolution of real-time and high-accuracy affine medical image registration algorithms is necessary for registration applications.Approach. In this paper, we propose a deep learning-based coarse-to-fine global and local feature fusion architecture for fast affine registration, and we use an unsupervised approach for end-to-end training. We use multiscale convolutional kernels as our elemental convolutional blocks to enhance feature extraction. Then, to learn the long-range spatial relationships of the features, we propose a new affine registration framework with weighted global positional attention that fuses global feature mapping and local feature mapping. Moreover, the fusion regressor is designed to generate the affine parameters.Main results. The additive fusion method can be adaptive to global mapping and local mapping, which improves affine registration accuracy without the center of mass initialization. In addition, the max pooling layer and the multiscale convolutional kernel coding module increase the ability of the model in affine registration.Significance. We validate the effectiveness of our method on the OASIS dataset with 414 3D MRI brain maps. Comprehensive results demonstrate that our method achieves state-of-the-art affine registration accuracy and very efficient runtimes.

Single station MILP scheduling in discrete and continuous time

AbstractThis article focuses on production planning in the metallurgical sector. This study undertakes a detailed comparative study of mixed-integer linear programming models using different time representations: continuous and discrete. The analysis shows that the continuous model consistently outperforms its discrete counterpart in all evaluated scenarios. The key difference between the continuous and discrete models is the continuous model’s ability to deliver better makespan results, achieving an improvement of up to 15% compared to the discrete model. This advantage holds even in complex environments with a high number of tasks and machines, where the continuous model consistently outperforms the discrete model by over 6% in the scenario with the highest number of tasks and machines. This preference extends beyond makespan considerations. The continuous model also maintains an edge in terms of runtime efficiency, achieving better times with a 99% improvement over the discrete model in all scenarios except one. These findings provide concrete evidence for the use of continuous models, which promise more effective production planning in analogous manufacturing domains.

A wall-time minimizing parallelization strategy for approximate Bayesian computation.

Approximate Bayesian Computation (ABC) is a widely applicable and popular approach to estimating unknown parameters of mechanistic models. As ABC analyses are computationally expensive, parallelization on high-performance infrastructure is often necessary. However, the existing parallelization strategies leave computing resources unused at times and thus do not optimally leverage them yet. We present look-ahead scheduling, a wall-time minimizing parallelization strategy for ABC Sequential Monte Carlo algorithms, which avoids idle times of computing units by preemptive sampling of subsequent generations. This allows to utilize all available resources. The strategy can be integrated with e.g. adaptive distance function and summary statistic selection schemes, which is essential in practice. Our key contribution is the theoretical assessment of the strategy of preemptive sampling and the proof of unbiasedness. Complementary, we provide an implementation and evaluate the strategy on different problems and numbers of parallel cores, showing speed-ups of typically 10-20% and up to 50% compared to the best established approach, with some variability. Thus, the proposed strategy allows to improve the cost and run-time efficiency of ABC methods on high-performance infrastructure.

Exploration of movie evaluation analysis and data preprocessing impact based on RNN technology

In order to provide valuable models for the film and television industry, this study aims to introduce recurrent neural network (RNN) techniques for effective movie evaluation analysis. Semantic analysis using machine learning is a very important means of extracting and understanding the meaning behind the text. The study evaluates different RNN techniques to identify the optimal neural network model. Data preprocessing includes tokenization and embedding, including dataset partitioning, tokenization process and word embedding techniques. The comparative analysis involves the predictive performance of simple RNN, Long Short-Term Memory Network (LSTM) and LSTM with attention. This study also explores the impact of including emoji and punctuation analysis in the data preprocessing process on semantic analysis. The results of the study show that preprocessing emoticons and punctuation improves accuracy, and LSTM with attention shows excellent performance. Notably, the study concludes that LSTM with attention performs well in terms of runtime efficiency, convergence speed, and accuracy compared to other models. The effect of punctuation and emoticons is that it will improve the accuracy. This study helps to improve the quality of the movie by constructing an effective analytical model thus.

Programming with ChatGPT: How far can we go?

Artificial intelligence (AI) has made remarkable strides, giving rise to the development of large language models such as ChatGPT. The chatbot has garnered significant attention from academia, industry, and the general public, marking the beginning of a new era in AI applications. This work explores how well ChatGPT can write source code. To this end, we performed a series of experiments to assess the extent to which ChatGPT is capable of solving general programming problems. Our objective is to assess ChatGPT’s capabilities in two different programming languages, namely C++ and Java, by providing it with a set of programming problem, encompassing various types and difficulty levels. We focus on evaluating ChatGPT’s performance in terms of code correctness, run-time efficiency, and memory usage. The experimental results show that, while ChatGPT is good at solving easy and medium programming problems written in C++ and Java, it encounters some difficulties with more complicated tasks in the two languages. Compared to code written by humans, the one generated by ChatGPT is of lower quality, with respect to runtime and memory usage.

A fast, scalable and versatile tool for analysis of single-cell omics data.

Single-cell omics technologies have revolutionized the study of gene regulation in complex tissues. A major computational challenge in analyzing these datasets is to project the large-scale and high-dimensional data into low-dimensional space while retaining the relative relationships between cells. This low dimension embedding is necessary to decompose cellular heterogeneity and reconstruct cell-type-specific gene regulatory programs. Traditional dimensionality reduction techniques, however, face challenges in computational efficiency and in comprehensively addressing cellular diversity across varied molecular modalities. Here we introduce a nonlinear dimensionality reduction algorithm, embodied in the Python package SnapATAC2, which not only achieves a more precise capture of single-cell omics data heterogeneities but also ensures efficient runtime and memory usage, scaling linearly with the number of cells. Our algorithm demonstrates exceptional performance, scalability and versatility across diverse single-cell omics datasets, including single-cell assay for transposase-accessible chromatin using sequencing, single-cell RNA sequencing, single-cell Hi-C and single-cell multi-omics datasets, underscoring its utility in advancing single-cell analysis.

WebQuorumChain: A web framework for quorum-based health care model learning

BackgroundInstitutions interested in collaborative machine learning to enhance healthcare may be deterred by privacy concerns. Decentralized federated learning is a privacy-preserving and security-robust tool to promote cross-institutional learning, however, such frameworks require complex setups and advanced technical expertise. Here, we aim to improve their utilization by offering an intuitive, user-friendly, and secure system that integrates both front-end and back-end functionalities. MethodWe develop WebQuorumChain, an integrated system built upon the QuorumChain schema. We test the system on a 2-site network using two publicly available health datasets and measure the average vertical and horizontal-ensemble AUCs per dataset across 30 trials, as well as the average execution time of the system. ResultsOur system achieved consistently high AUCs for each dataset (0.94–0.96), with reasonable total execution times ranging from 5 to 20 min, inclusive of modeling and all other system overheads. The front-end displays event logs generated from back-end layers in real time, in sync with the progress of the underlying algorithm. ConclusionsWe develop a web-based system that supplies users with visual tools to configure the federated learning network, manage training sessions, and inspect the learning process. WebQuorumChain helps schedule and monitor low-level processes without violating the fundamental security promises of cross-institutional decentralized machine learning. The system also maintains predictive accuracy and runtime efficiency in the presence of additional layers. WebQuorumChain will help promote meaningful collaboration among healthcare institutions, who can retain full control of their data privacy while contributing to data-driven discoveries.

Semantic Image Segmentation by Dynamic Discriminative Prototypes

Semantic segmentation achieves significant success through large-scale training data. Meanwhile, few-shot semantic segmentation was proposed to segment image regions of novel classes through few labeled testing data. However, it ignores classes previously learned from training data. This paper proposes a new segmentation framework called segmentation by dynamic prototype (SDP), which can simultaneously segment the image regions of base classes learned from many training data and novel classes learned from a few testing data. SDP performs segmentation by searching for the nearest prototype of each pixel's features. Different prototypes are representative features for different classes. In testing, SDP dynamically constructs novel prototypes according to support images while maintaining base prototypes learned from training images. The main challenge of SDP is how to achieve intra-image compactness, intra-class compactness, and inter-class separability simultaneously in the feature space. To tackle these challenges, we first introduce a discriminative pixelwise feature and prototype training method to improve the above three types of feature discriminability. We then introduce a mask refinement process in testing, which refines support images' masks to extract more separable novel prototypes. In addition, we introduce a prototype adaptation process in testing, which allows all prototypes to adapt to query images to reduce the prototypes' intra-class variances. Our approach achieves state-of-the-art results for novel class segmentation against existing few-shot semantic segmentation methods on PASCAL-5 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i</sup> and COCO-20 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i</sup> benchmarks in combination with superior runtime efficiency. In addition, our method can maintain strong results in the segmentation of base classes.

Effect of reservoir heterogeneity on well placement prediction in CO2-EOR projects using machine learning surrogate models: Benchmarking of boosting-based algorithms

Rising Carbon Dioxide (CO2) levels from human activities are driving climate change. Carbon capture and storage (CCS) during enhanced oil recovery (EOR) in underground reservoirs offer both environmental and economic benefits. This method boosts oil production, cuts greenhouse gas emissions, and supports sustainable energy. Precise well placement in CO2-EOR is a crucial task for effective oil displacement, but traditional reservoir simulators are costly. This study explores and compares boosting algorithms, as fast surrogate models, to achieve accurate well placement during CO2-EOR in light oil carbonate reservoirs. The research considers various reservoir scenarios with different geological heterogeneity levels (i.e., homogeneous, moderately heterogeneous, and highly heterogeneous reservoirs). Various parameters, such as injection and production well locations, the distance between production and injection wells in an inverted five-spot pattern, pattern angle, and injection and production rates are explored using a compositional reservoir simulator to assess their impact on the well placement problem. A comprehensive analysis of various boosting algorithms, including AdaBoost, CatBoost, Gradient Boosting, LightGBM, and XGBoost is performed using the simulated dataset to assess their efficacy. The results demonstrate that LightGBM outperformed the other algorithms with the lowest Mean Absolute Error and Root Mean Square Error of 115.3 × 106 $ and 188.2 × 106 $, respectively. Additionally, it demonstrates exceptional speed, averaging 3 to 8 times faster than other boosting algorithms in the three reservoir scenarios. This superior performance coupled with its efficient runtime makes LightGBM the ideal choice for the study objectives. Moreover, the mass balance approach highlights the significant CO2 storage efficiency, emphasizing the effectiveness of CO2-EOR in storing CO2 in underground heterogeneous reservoirs.

DGC: Training Dynamic Graphs with Spatio-Temporal Non-Uniformity using Graph Partitioning by Chunks

Dynamic Graph Neural Network (DGNN) has shown a strong capability of learning dynamic graphs by exploiting both spatial and temporal features. Although DGNN has recently received considerable attention by AI community and various DGNN models have been proposed, building a distributed system for efficient DGNN training is still challenging. It has been well recognized that how to partition the dynamic graph and assign workloads to multiple GPUs plays a critical role in training acceleration. Existing works partition a dynamic graph into snapshots or temporal sequences, which only work well when the graph has uniform spatio-temporal structures. However, dynamic graphs in practice are not uniformly structured, with some snapshots being very dense while others are sparse. To address this issue, we propose DGC, a distributed DGNN training system that achieves a 1.25× - 7.52× speedup over the state-of-the-art in our testbed. DGC's success stems from a new graph partitioning method that partitions dynamic graphs into chunks, which are essentially subgraphs with modest training workloads and few inter connections. This partitioning algorithm is based on graph coarsening, which can run very fast on large graphs. In addition, DGC has a highly efficient run-time, powered by the proposed chunk fusion and adaptive stale aggregation techniques. Extensive experimental results on 3 typical DGNN models and 4 popular dynamic graph datasets are presented to show the effectiveness of DGC.

Multi-level deep domain adaptive adversarial model based on tensor-train decomposition for industrial time series forecasting

The polyester industry is a complex process industry, building a time series prediction model for new production lines or equipment with new sensors can be challenging due to a lack of historical data. The time-series data collected from sensors cross-production-line often exhibit varying distributions. Current domain adaptation (DA) approaches in data-driven time series forecasting primarily concentrate on adjusting either the features or the models, neglecting the intricacies of industrial time series data. Furthermore, constructing deep neural networks for industrial data necessitates substantial computational resources and runtime due to their large and high-dimensional nature. In order to tackle these obstacles, we propose a novel Multi-level deep domain adaptive adversarial model based on tensor-train decomposition (TT-MDAM). Our model aims to strike a dynamic balance between prediction accuracy and runtime efficiency. By integrating multiple perspectives at the feature, trend, and model levels, we leverage DA to enhance the prediction accuracy of our model in the target domain. Additionally, by analyzing the weight matrix of the neural network, we generate a low-rank model to improve operational efficiency. The application of the proposed TT-MDAM approach to both the three-phase flow facility process (TPFF) dataset and a real-world polyester esterification process dataset reveals promising results, outperforming state-of-the-art methodologies in terms of prediction performance. The results indicate that the approach provides a viable solution for building time series prediction models in industrial processes with new equipment or production lines.

Multilevel thresholding image segmentation algorithm based on Mumford–Shah model

Abstract Image segmentation is one of the important tasks of computer vision and computer image processing, and the purpose of image segmentation is to achieve the extraction and recognition of the target image region. The classical Mumford–Shah (MSh) image segmentation model is used to achieve the segmentation of images. With the goal to get the best segmentation effect on images by minimizing the MSh energy generalization function, a level set strategy is developed, and a model with global information infinite curve evolution is utilized. However, considering the low efficiency of this model for processing level set curves and the general quality of image segmentation. A multi-layer threshold search scheme is proposed to achieve rapid convergence of the target image level set curve. The experimental results showed that the multi-level thresholding image segmentation algorithm based on the MSh model can significantly improve the segmentation effect of images and reduce the segmentation time. The suggested MSK method outperforms the MPO algorithm, SSA algorithm, and EMO algorithm in the picture segmentation convergence time test, respectively, in terms of runtime efficiency by 356, 289, and 71%. Additionally, it performs superbly in both threshold searches and picture quality tests. The research topic has significant reference value for the study of contemporary computer vision imaging technologies.

Pipelines for automating compliance-based elimination and extension (PACE2): a systematic framework for high-throughput biomolecular materials simulation workflows

The formation of biomolecular materials via dynamical interfacial processes, such as self-assembly and fusion, for diverse compositions and external conditions can be efficiently probed using ensemble Molecular Dynamics (MD). However, this approach requires many simulations when investigating a large composition phase space. In addition, there is difficulty in predicting whether each simulation will yield biomolecular materials with the desired properties or outcomes and how long each simulation will run. These difficulties can be overcome by rules-based management systems, including intermittent inspection, variable sampling, and premature termination or extension of the individual MD simulations. Automating such a management system can significantly improve runtime efficiency and reduce the burden of organizing large ensembles of MD simulations. To this end, a computational framework, the Pipelines for Automating Compliance-based Elimination and Extension (PACE2), is proposed for high-throughput ensemble biomolecular materials simulations. The PACE2 framework encompasses Candidate pipelines, where each pipeline includes temporally separated simulation and analysis tasks. When a MD simulation is completed, an analysis task is triggered, which evaluates the MD trajectory for compliance. Compliant simulations are extended to the next MD phase with a suitable sample rate to allow additional, detailed analysis. Non-compliant simulations are eliminated, and their computational resources are reallocated or released. The framework is designed to run on local desktop computers and high-performance computing resources. Preliminary scientific results enabled by the use of PACE2 framework are presented, which demonstrate its potential and validates its function. In the future, the framework will be extended to address generalized workflows and investigate composition-structure-property relations for other classes of materials.

Parallelization and optimization for a pin-by-pin whole core thermal–hydraulic analysis code

With increasingly developing high-performance computing (HPC) technology, pin-by-pin thermal–hydraulic simulations are now feasible. The in-house subchannel analysis code KMC-FBc has been extended to perform pin-by-pin calculations. To expedite the whole-core thermal–hydraulic analysis process, various parallelization and optimization strategies are proposed and implemented at the single-assembly and multi-assembly levels, including acceleration methods for solving large sparse matrix equation systems, memory and file input/output (I/O) optimization, hybrid MPI/OpenMP-based whole-core acceleration method and so on. Parameters for evaluating the parallel performance, such as the running time, speedup, and parallel efficiency, are tested on an HPC cluster. Results show that the maximum acceleration ratio for the whole-core problem with a mesh size in the tens of millions can reach approximately 120, with a parallel efficiency of around 32%, which demonstrates the rationality and effectiveness of the parallel algorithm and optimization methods and lays the foundation for future high-resolution and high-fidelity thermal–hydraulic analysis.

A Fast FPGA Connection Router Using Prerouting-Based Parallel Local Routing Algorithm

Routing is one of the most time-consuming steps in the FPGA design process. Even if unceasing efforts have been made to accelerate FPGA routing, the existing work seldom pays attention to the underlying FPGA connection router. In this paper, we present a fast FPGA connection router called PRoute which implements a novel pre-routing based parallel local routing algorithm. Basically, PRoute pre-computes the potential routing solutions for various connection patterns on FPGAs, which can be directly used in the later practical routing. On the whole, PRoute is composed of A-star maze expansion and parallel local search. In the first part, PRoute gradually expands the maze wavefront towards the lowest-cost node to search for the target sink. For a wire-type node, PRoute invokes a fast parallel local search instead taking advantage of the pre-routing results, and hence the time-expensive maze expansion can be reduced. Particularly, it allows PRoute to call one another between A-star maze expansion and parallel local search. This enables the runtime efficiency of PRoute while ensuring its global search ability. In addition, we put forward an engineering improvement to further speed up PRoute by avoiding the exploration of block output pins. To our best knowledge, this work is the first to apply the idea of pre-routing for FPGAs. Experimental results show that PRoute achieves speedups of 1.8×, 2.4×, 3.2×, 4.1× and 5.1× with 1, 4, 8, 16 and 32 threads relative to the baseline VPR’s connection router respectively, without degrading the quality of results.

Real-Time 3D Object Detection on Crowded Pedestrians.

In the field of autonomous driving, object detection under point clouds is indispensable for environmental perception. In order to achieve the goal of reducing blind spots in perception, many autonomous driving schemes have added low-cost blind-filling LiDAR on the side of the vehicle. Unlike point cloud target detection based on high-performance LiDAR, the blind-filling LiDARs have low vertical angular resolution and are mounted on the side of the vehicle, resulting in easily mixed point clouds of pedestrian targets in close proximity to each other. These characteristics are harmful for target detection. Currently, many research works focus on target detection under high-density LiDAR. These methods cannot effectively deal with the high sparsity of the point clouds, and the recall and detection accuracy of crowded pedestrian targets tend to be low. To overcome these problems, we propose a real-time detection model for crowded pedestrian targets, namely RTCP. To improve computational efficiency, we utilize an attention-based point sampling method to reduce the redundancy of the point clouds, then we obtain new feature tensors by the quantization of the point cloud space and neighborhood fusion in polar coordinates. In order to make it easier for the model to focus on the center position of the target, we propose an object alignment attention module (OAA) for position alignment, and we utilize an additional branch of the targets' location occupied heatmap to guide the training of the OAA module. These methods improve the model's robustness against the occlusion of crowded pedestrian targets. Finally, we evaluate the detector on KITTI, JRDB, and our own blind-filling LiDAR dataset, and our algorithm achieved the best trade-off of detection accuracy against runtime efficiency.

An advanced check‐dam sedimentation module: Laboratory validation and implementation in a distributed sediment yield model for field application

AbstractCheck‐dams are structures used extensively around the world for soil and water conservation. However, existing models for check‐dams are unable to simulate the Sediment Trap Efficiency (STE) at the catchment scale. A numerical model was developed to simulate the SEdiment Deposition upstream of Check‐Dams (SEDCD) and integrated into a distributed sediment yield model, the Digital Yellow River Model (DYRIM). Two versions of the SEDCD model were evaluated: the SV version which used the Saint–Venant equation to simulate the unsteady flow, and the BW version, which used a modified backwater equation (based on a quasi‐steady approximation) to improve computational efficiency. Sediment deposition and the associated bed profile adjustment were simulated with the sediment conservation equation and the non‐equilibrium suspended sediment transport equation. The SEDCD model was first validated in the laboratory using experimental data from a scale‐down check‐dam. The bed profiles predicted using both versions of the SEDCD model showed good agreement with the observations, with NSE values over 0.9 in most profiles. When integrated into the DYRIM and applied to the Xiaoli River Basin (818 km2) on the Loess Plateau, which has 183 active check‐dams, the SEDCD‐DYRIM combination predicted the STE for an extreme rainstorm event in 2017 with good accuracy and high computational efficiency. The SEDCD‐BW‐DYRIM simulated the hourly discharge and sediment concentration with high accuracy (NSE values of 0.79 and 0.71, respectively) and provided single‐event STEs (R2 value of 0.99) comparable to those of the SEDCD‐SV model, with an approximately 30 times faster runtime efficiency than the SEDCD‐SV model. The SEDCD‐BW model is a powerful and efficient tool to assess the effect of check dams on sediment dynamics at the catchment scale.