Running Time Research Articles

A new efficient decoder of linear block codes based on ensemble learning models: case of boosting

<p>Error-correcting codes are used to partially or completely correct errors as much as possible, while ensuring high transmission speeds. Several Machine Learning (ML) models such as Logistic Regression and Decision tree have been applied to correct transmission errors. Among the most powerful ML techniques are aggregation methods which have yielded to excellent results in many areas of research. It is this excellence that has prompted us to consider their application for the hard decoding problem. In this sense, we have successfully designed, tested and validated our proposed EL-BoostDec decoder (hard decision decoder based on Ensemble Learning - Boosting technique) which is based on computing of the syndrome of the received word and on using Ensemble Learning techniques to find the corresponding corrigible error. The obtained results with EL-BoostDec are very encouraging in terms of the binary error rate (BER) that it offers. Practically EL-BoostDec has succeed to correct 100% of errors that have weights less than or equal to the correction capacity of studied codes. The comparison of EL-BoostDec with many competitors proves its power. A study of parameters which impact on EL-BoostDec performances has been established to obtain a good BER with minimum run time complexity.</p>

Face recognition with occluded face using improve intersection over union of region proposal network on Mask region convolutional neural network

Face recognition entails detecting and identifying facial attributes. Mask region convolutional neural network (R-CNN) method is a prominent approach, while prior research predominantly delved into refining loss functions and perfecting object and face detection, recognizing, and identifying faces using imperfect data remained relatively unexplored. This study focuses on an occluded dataset comprising Indonesian faces, wherein 'occluded' denotes facial data that lacks complete visibility-encompassing instances where objects obscure faces or are partially cropped. This investigation involves a deliberate experiment that tailors the intersection over union (IoU) of the region proposal network (RPN) to suit the nuances of occluded Indonesian faces, thereby augmenting accuracy in recognition and segmentation tasks. The innovation IoU in the strategic utilization of Anchors, which involves the exclusion of anchors falling beyond the image borders to optimize computational efficiency. The outcomes of this research are striking; it showcases a remarkable 14.75%, 10.9%, and 12.97% surge based on mean average precision (mAP), mean average recall (mAR), and F1-Scores compared to the conventional Mask R-CNN approach. Notably, our proposed model elevates the average accuracy by 10% to 15% and decreases running time by 21%, a noteworthy enhancement compared to the preceding model. This progress is substantiated by validation utilizing 300 instances dataset, reinforcing the robustness of our approach.

Effect of 6-Week Sprint Training on Long-Distance Running Performance in Highly Trained Runners.

Long-distance running performance has been reported to be associated with sprint performance in highly trained distance runners. Therefore, we hypothesized that sprint training could enhance distance running and sprint performance in long-distance runners. This study examined the effect of 6-week sprint training on long-distance running and sprint performance in highly trained distance runners. Nineteen college runners were divided into control (n = 8) and training (n = 11) groups. Participants in the training group performed 12 sprint training sessions in 6weeks, while those in the control group performed 12 distance training sessions. Before and after the interventions, maximal oxygen uptake (V˙O2max), O2 cost during submaximal running (290m·min-1 and 310m·min-1 of running velocity), and time to exhaustion (starting at 290m·min-1 and increased 10m·min-1 every minute) were assessed on a treadmill. Additionally, the 100-m and 400-m sprinting times and 3000-m running time were determined on an all-weather track. In the control group, no measurements significantly changed after the intervention. In the training group, the time to exhaustion, 100-m and 400-m sprinting times, and 3000-m running time improved significantly, while V˙O2max and O2 cost did not change. These results showed that 6-week sprint training improved both sprint and long-distance running performance in highly trained distance runners without a change in aerobic capacity. Improvement in the time to exhaustion without a change in V˙O2max suggests that the enhancement of long-distance running performance could be attributable to improved anaerobic capacity.

A Comprehensive Review on Advantages of Ultra Performance Liquid Chromatography (UPLC) for Analytical Method Development, Validation, and Stability Studies

Ultra Performance Liquid Chromatography (UPLC) represents a paradigm shift in analytical separation science, offering enhanced speed, sensitivity, and resolution while retaining the foundational principles of traditional liquid chromatography. This comprehensive review explores the principles, instrumentation, advantages, and applications of UPLC in analytical method development, validation, and stability studies, with a focus on pharmaceutical analysis. UPLC leverages finer particles and intensified flow rates to achieve unprecedented performance gains, facilitating rapid and precise separations without compromising analytical rigor. The discussion delves into the intricate instrumentation involved, including sample injection, column selection, and detector operation, highlighting the critical role of each component in ensuring optimal analytical performance. Advantages of UPLC include drastically reduced run times, streamlined method optimization, and increased sample throughput, while limitations such as higher back pressures and limited column regenerability are also discussed. Furthermore, the review outlines the diverse applications of UPLC across pharmaceutical, environmental, food and beverage, clinical, forensic, bioanalytical, and polymer analysis domains, underscoring its versatility and significance in advancing research, development, and quality control in various industries. In summary, UPLC represents a transformative tool in analytical chemistry, offering unparalleled capabilities for expedited and accurate analysis of complex sample matrices, ultimately driving innovation and scientific progress. Key Words: Ultra Performance Liquid Chromatography (UPLC), analytical method development, validation, stability studies, pharmaceutical analysis, instrumentation, advantages, applications, sample injection, column chemistry, detector technology, back pressures, method optimization, sample throughput, environmental analysis, food and beverage analysis, clinical diagnostics, forensic analysis, bioanalytical research, polymer analysis, analytical chemistry.

Accurate wire rope defect MFL detection using improved Hilbert transform and LSTM neural network

ABSTRACT Wire rope is playing a significant role in various engineering fields. Although a great number of non-destructive testing methods are applied to wire rope defect detection, the results are greatly influenced by the signal noises and the quantitative defect detection of wire rope is consequently limited. Therefore, a new accurate and quantitative wire rope defect magnetic flux leakage (MFL) recognition method based on improved Hilbert transform and long short-term memory (LSTM) neural network is proposed. The theoretical background of the proposed signal processing models and preliminary results, as well as the influence of the main model parameters are analyzed first. Then, different wire rope defect classification principles through LSTM neural network and performance evaluation based on the improved signal processing and machine learning combined method are presented. Finally, the defect inspection and classification comparison results between the proposed method and other deep learning neural networks and pattern recognition strategies are given, which manifest that the proposed method has higher classification accuracy and shorter running time for five different wire rope defects under various servicing conditions, and is promising in accurate defect inspection for multi-scenario wire rope. Additionally, the main conclusions, disadvantages of the proposed method are summarized.

STABILITY BASED HPLC METHOD FOR CYCLOPHOSPHAMIDE RELATED SUBSTANCES IN FINISHED DRUG PRODUCTS: DEVELOPMENT AND VALIDATION

Objective: The current study aimed to develop a simple, sensitive, and precise high-performance liquid chromatographic (HPLC) method for estimating cyclophosphamide and its related substances, as well as to implement the developed method in a capsule product. Methods: Method development was performed using various solvent and buffer-solvent ratios at different flow rates for better resolution and to decrease the run time. The developed method was validated in accordance with the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) guidelines. The developed method was implemented to estimate the amount of cyclophosphamide and its related substances. Results: Chromatographical conditions were optimised, and the best chromatographical conditions with adequate resolution for cyclophosphamide and its related substances were achieved using the enable X-Bridge C18 column, using a mobile phase combination of phosphate buffer pH 7.0, water, methanol, and acetonitrile at a flow rate of 0.5 ml/min. The detection was monitored at a wavelength of 200 nm. The developed method was validated for system suitability, specificity, limit of detection (LOD), limit of quantitation (LOQ), linearity, precision, accuracy and robustness. The results indicate that the method was sensitive and could detect and quantify lower levels of cyclophosphamide and its related substances. The linearity range was from LOQ to 150 %, and a correlation coefficient ranging from 0.9987 to 0.9999 indicates that at this concentration range, substances were highly linear. This was further supported by satisfactory forced degradation studies. Conclusion: The developed analytical method is simple, precise, and reproducible and thus can be used for the stability-indicating analysis of cyclophosphamide and its related substances in pharmaceutical formulations.

Detail-preserving Joint Image Upsampling

Image operators can be instrumental to computational imaging and photography. However, many of them are computationally intensive. In this paper, we propose an effective yet efficient joint upsampling method to accelerate various image operators. We show that edge-preserving filtering can be facilitated with a downsampling-and-upsampling process. Moreover, when the extent of smoothing is mild, the process is detail-preserving, i.e., the fine details lost in the low-resolution (LR) images can be accurately restored in the high-resolution (HR) images. Given an HR input and an LR output of an operator, we downsample the HR input and calculate its affinities to the HR input. By applying the affinities to the LR output, we promote its resolution. Due to the strong detail-preserving property, the HR output derived in the previous step may exhibit aliasing artifacts around the salient edges. We further refine it based on the linear relations in a small neighborhood to rid the artifacts. Experiments on various image operators show that our method achieves superior quality over the state-of-the-art joint upsampling methods. Furthermore, the running time of our method is linear to the number of pixels. Our naive implementation derives 1080P images in real time (24 fps) on an NVIDIA GTX 3070 GPU.

Transfer learning enables identification of multiple types of RNA modifications using nanopore direct RNA sequencing

Nanopore direct RNA sequencing (DRS) has emerged as a powerful tool for RNA modification identification. However, concurrently detecting multiple types of modifications in a single DRS sample remains a challenge. Here, we develop TandemMod, a transferable deep learning framework capable of detecting multiple types of RNA modifications in single DRS data. To train high-performance TandemMod models, we generate in vitro epitranscriptome datasets from cDNA libraries, containing thousands of transcripts labeled with various types of RNA modifications. We validate the performance of TandemMod on both in vitro transcripts and in vivo human cell lines, confirming its high accuracy for profiling m6A and m5C modification sites. Furthermore, we perform transfer learning for identifying other modifications such as m7G, Ψ, and inosine, significantly reducing training data size and running time without compromising performance. Finally, we apply TandemMod to identify 3 types of RNA modifications in rice grown in different environments, demonstrating its applicability across species and conditions. In summary, we provide a resource with ground-truth labels that can serve as benchmark datasets for nanopore-based modification identification methods, and TandemMod for identifying diverse RNA modifications using a single DRS sample.

A two‐dimensional, reach‐scale implementation of space‐time image velocimetry (STIV) and comparison to particle image velocimetry (PIV)

AbstractImage‐based algorithms have become a powerful tool for estimating flow velocities in rivers. In this study, we generalize the space‐time image velocimetry (STIV) framework for reach‐scale application rather than along a cross section. The new algorithm provides information on both the magnitude and orientation of velocity vectors, and we refer to the algorithm as two‐dimensional STIV, or 2D‐STIV. The workflow involves setting up a grid, using centreline tangent vectors as initial estimates of flow direction, and then extracting space‐time images (STIs) along search lines radiating from each grid node. The autocorrelation function is used to infer the inclination of streak lines present in STIs, which represents the advection of water surface features. Information on flow direction is obtained by evaluating various candidate search lines and identifying that which yields the highest velocity. This search can be performed exhaustively or via optimization. We applied the new 2D‐STIV algorithm to three test cases, one simulated data set and two natural channels, and compared image‐derived velocities to modelled or measured values. We also applied two established particle image velocimetry (PIV) algorithms to the same data sets. 2D‐STIV performed as well as the two PIV algorithms for simulated images. For a natural river with distinct water surface features, 2D‐STIV was effective for much of the channel but also led to a more patchy, irregular velocity field than the two PIV algorithms. For a site lacking obvious surface features, exhaustive 2D‐STIV led to velocity estimates uncorrelated with field data while the optimization‐based version produced erratic flow directions. 2D‐STIV also required greater image sequence durations, higher frame rates, and generally longer computational run times. Overall, ensemble PIV was the most reliable algorithm.

Expanding the wrench feasible workspace of quadrotor-based mobile cable-driven parallel manipulators using multi-objective optimization and machine learning

Quadrotor-based Reconfigurable Cable-Driven Parallel Manipulators (QRCDPMs) have previously been introduced as having significant potential in extending the feasible workspace of Cable-Driven Parallel Manipulators (CDPMs). However, such configurations have limitations in executing positive Z-direction trajectories resulting in non-optimal workspace and performance. This study proposes a novel Quadrotor-based Enhanced Reconfiguration Cable-Driven Parallel Manipulator (QERCDPM) that significantly enhances the Wrench Feasible Workspace (WFW). This allows for extended safe platform operation along all XYZ directions by adjusting the bottom cable exit point. Simulation results indicate that the proposed setup achieves a 96.39 % ratio of safe workspace to available workspace in various cable exit points configurations. Tension Distribution analysis for different trajectories demonstrates the proposed setup's superior cable tension management, increased flexibility and reachability. A Random Forest based Particle Swarm Optimization - Modified Frequency Bat Algorithm (RFPSOMFB) optimization algorithm is proposed, which outperforms existing methods in terms of run time and solution quality. The findings hold significance for applications requiring safe and efficient trajectory execution in dynamic and constrained environments, contributing to the evolution of multi-robot based intelligent and distributed systems.

Stripe-schedule Aware Repair in Erasure-coded Clusters with Heterogeneous Star Networks

More and more storage systems use erasure code to tolerate faults. It takes pieces of data blocks as input and encodes a small number of parity blocks as output, where these blocks form a stripe. When reconsidering the recovery problem in the multi-stripe level and heterogeneous network clusters, quickly generating an efficient multi-stripe recovery solution that reduces recovery time remains a challenging and time-consuming task. Previous works either use a greedy algorithm that may fall into the local optimal and have low recovery performance or a meta-heuristic algorithm with a long running time and low solution generation efficiency. In this paper, we propose a Stripe-schedule Aware Repair (SARepair) technique for multi-stripe recovery in heterogeneous erasure-coded clusters based on RS code. By carefully examining the metadata of blocks, SARepair intelligently adjusts the recovery solution for each stripe and obtains another multi-stripe solution with less recovery time in a computationally efficient manner. It then tolerates worse solutions to overcome the local optimal and uses a rollback mechanism to adjust search regions to reduce recovery time further. Moreover, instead of reading blocks sequentially from each node, SARepair also selectively schedules the reading order for each block to reduce the memory overhead. We extend SARepair to address the full-node recovery and adapt to the LRC code. We prototype SARepair and show via both simulations and Amazon EC2 experiments that the recovery performance can be improved by up to 59.97% over a state-of-the-art recovery approach while keeping running time and memory overhead low.

Optimizing time and cost in construction projects with a hybridized multi-verse optimizer and opposition-based learning

PurposeFor successful management of construction projects, a precise analysis of the balance between time and cost is imperative to attain the most effective results. The aim of this study is to present an innovative approach tailored to tackle the challenges posed by time-cost trade-off (TCTO) problems. This objective is achieved through the integration of the multi-verse optimizer (MVO) with opposition-based learning (OBL), thereby introducing a groundbreaking methodology in the field.Design/methodology/approachThe paper aims to develop a new hybrid meta-heuristic algorithm. This is achieved by integrating the MVO with OBL, thereby forming the iMVO algorithm. The integration enhances the optimization capabilities of the algorithm, notably in terms of exploration and exploitation. Consequently, this results in expedited convergence and yields more accurate solutions. The efficacy of the iMVO algorithm will be evaluated through its application to four different TCTO problems. These problems vary in scale – small, medium and large – and include real-life case studies that possess complex relationships.FindingsThe efficacy of the proposed methodology is evaluated by examining TCTO problems, encompassing 18, 29, 69 and 290 activities, respectively. Results indicate that the iMVO provides competitive solutions for TCTO problems in construction projects. It is observed that the algorithm surpasses previous algorithms in terms of both mean deviation percentage (MD) and average running time (ART).Originality/valueThis research represents a significant advancement in the field of meta-heuristic algorithms, particularly in their application to managing TCTO in construction projects. It is noteworthy for being among the few studies that integrate the MVO with OBL for the management of TCTO in construction projects characterized by complex relationships.

Drone scheduling in a hexagon‐based delivery network with multiple fulfillment centers

AbstractAs an emerging technology, drones have great potential in small parcel delivery, reducing operational costs, and improving efficiency. However, battery technologies restrict the endurance of drones. Inspired by the cellular networks used in mobile communications, a hexagon‐based delivery network (HBDN) is proposed to overcome such restrictions. Battery‐swapping stations are incorporated into the HBDN to expand the delivery service range of drones. Based on the proposed HBDN, the multidrone scheduling problem in multiple fulfillment centers (FCs) is addressed. The objective is to minimize the makespan. A mixed‐integer programming model is constructed for the problem. A memetic algorithm with hill climbing (MAHC) is developed to address the problem. A strategy is proposed to prevent inbreeding in the crossover procedure. A local search based on hill climbing is combined with the mutation operation. A lower bound of the problem is proposed to measure the performance of MAHC. Furthermore, load balancing and nearest‐priority strategies are proposed to select FCs. Comparative experiments show that MAHC performs competitively in terms of running time and solution quality.

Local Path Planner for Mobile Robot Considering Future Positions of Obstacles

Local path planning is a necessary ability for mobile robot navigation, but existing planners are not sufficiently effective at dynamic obstacle avoidance. In this article, an improved timed elastic band (TEB) planner based on the requirements of mobile robot navigation in dynamic environments is proposed. The dynamic obstacle velocities and TEB poses are fully integrated through two-dimensional (2D) lidar and multi-obstacle tracking. First, background point filtering and clustering are performed on the lidar points to obtain obstacle clusters. Then, we calculate the data association matrix of the obstacle clusters of the current and previous frame so that the clusters can be matched. Thirdly, a Kalman filter is adopted to track clusters and obtain the optimal estimates of their velocities. Finally, the TEB poses and obstacle velocities are associated: we predict the obstacle position corresponding to the TEB pose through the detected obstacle velocity and add this constraint to the corresponding TEB pose vertex. Then, a pose sequence considering the future positions of obstacles is obtained through a graph optimization algorithm. Compared with the original TEB, our method reduces the total running time by 22.87%, reduces the running distance by 19.23%, and increases the success rate by 21.05%. Simulations and experiments indicate that the improved TEB enables robots to efficiently avoid dynamic obstacles and reach the goal as quickly as possible.

Firm value adjustment speed through financial friction in the presence of earnings management and productivity growth: evidence from emerging economies

This study investigates the role of financial frictions on firm value within the framework of earnings management, including the impact of productivity growth. In contrast to prior studies, the present study employed an autoregressive model to examine the temporal dynamics of the variables to determine their short-term and long-term connection patterns. The results of the study indicate a negative association between financial frictions and firm value. Accrual earnings management, a practice employed by organizations to enhance their profit margins, serves as a mediator between financial frictions and firm value. This mediation of earnings management alleviates the adverse impact of financial frictions. The enhancement of productivity growth amplifies the conditional, indirect influence of earnings management. Moreover, this study reveals that financial frictions have a significant influence in the short-term, leading to overestimation of factor loadings. However, this impact stabilizes over time in the long run. Financial market frictions have the most prominent impact on firm value compared to the other two forms of frictions, namely, macroeconomic frictions and microeconomic frictions. Larger firms are more inclined to attain higher firm value than smaller enterprises. Managers can enhance firm value by exerting control over the influence of financial frictions in the economy through earnings management. The effectiveness of this strategy is contingent upon the level of productivity growth.

Analyzing the impact of timetable elements on railway line capacity

Train timetabling poses inherent challenges, prompting the need to enhance existing schedules by extracting valuable insights from the current timetable structure. By such philosophy, this paper studies the impact of timetable elements on the consumed capacity. These elements primarily encompass train operation parameters, including running times and stop plans. Their impact is defined as the deviation of consumed capacity relative to their variations. We initially establish a link between consumed capacity and timetable elements. This relationship is articulated as the signed sum of timetable elements along a designated ”critical path”. Then, the limited impact of any timetable element is clarified, namely changing an element can impact the consumed capacity with its neighbor trains in a combinatorial way. With this knowledge, we analyze the impact of a single element, using stop plans for example. This result is then generalized into analyzing the impact of several dependent and independent stop plans. The findings on capacity calculation and impact analysis of a single element are tested through real-world numerical computations and then extended to analyzing various capacity factors, such as average speed and heterogeneity.

Training mode of university dance performers based on sequential quadratic programming algorithm

Dance performance is an art form, which needs to cultivate students’ dance skills, artistic accomplishment and stage performance ability. Sequential quadratic programming algorithm is an optimization algorithm that can be used to solve complex optimization problems. In this paper, Sequential Quadratic Programming (SQP) is applied to explore the training mode of dance performers in colleges to help dance performers develop the optimal training plan and program. Aiming at the problems existing in the training mode of dance performance talents in colleges, this paper put forward an optimization method based on SQP algorithm, and implemented its optimization scheme in actual colleges. In the planning of dance performance talent training mode, particle swarm optimization (PSO) is used to optimize SQP algorithm, so that it can have higher planning efficiency. This paper studied the goal and index system of the training of dance performance talents in colleges, generated a personalized training program, and further improved the scientific and practical effectiveness of the training mode. This paper investigated the current situation of dance performance talent training in several dance schools in a certain province of China. The survey data include practical curriculum planning, teachers’ teaching philosophy and teaching content. Combined with SQP algorithm, the teaching and training program is optimized. After evaluation, it can be concluded that the SQP algorithm optimized by PSO shows good stability and accuracy. It can calculate the optimal solution of the cultivation scheme, and when calculating the optimal solution, the running time of the Central Processing Unit (CPU) was only 5.6 s, which can further improve the efficiency of the planning. Finally, through the satisfaction and resource utilization test, it can be found that the number of people who are very satisfied with the teaching content of the dance performance talent training program optimized by SQP increased from 38.4% to 52.4%. After optimization, the average utilization rate of teaching resources reached 88.1%. It can be concluded that SQP algorithm can provide scientific basis for dance education institutions to improve the training mode of dance talents. This can help dance education institutions better improve the training mode, and improve the overall quality and dance skills of dance students.

Hyperspectral image classification based on superpixel merging and broad learning system

AbstractMost spectral–spatial classification methods for hyperspectral images (HSIs) can achieve satisfactory classification results. However, the common problem faced with these approaches is the need for a long training time and sufficient training samples. To address this issue, this study proposes an effective spectral–spatial HSI classification method based on superpixel merging, superpixel smoothing and broad learning system (SMS‐BLS). The newly introduced parameter‐free superpixel merging technique based on local modularity not only enhances the role of local spatial information in classification, but also maintains class boundary information as much as possible. In addition, the spectral and spatial information of HSIs is further fused during the superpixel smoothing process. As a result, with limited training samples, using merged and smoothed superpixels instead of pixels as input to the broad learning system significantly improves its classification performance. Moreover, the merged superpixels weaken the dependence of the classification results on the superpixel segmentation scale. The effectiveness of the proposed method was validated on three HSI benchmarks, namely Indian Pines, Pavia University and Salinas. Experimental and comparative results show the superiority of the method to other state‐of‐the‐art approaches in terms of overall accuracy and running time.

Event driven high speed data acquisition with IEEE 1588 synchronization for long pulse operations of Indian Test Facility for ITER DNB

The Indian Test Facility (INTF) is a negative Hydrogen ion based 100 kV, 60A, 5 Hz modulated NBI system having duty cycle of 3 s ON time to 20 s OFF time. Prime objective of the facility is to characterize ITER Diagnostic Neutral Beam (DNB) with full specifications, prior to shipment and installation in ITER. In order to fully characterize ITER DNB, a matured Data Acquisition and Control system (DACS) is needed which will be responsible for supervisory control of the facility and for acquiring and monitoring signals at run time. Apart from conventional DACS, high speed data acquisition is needed for diagnostics and other signals with sampling rate requirements of 100 MS/s. For acquiring data from such systems it is imperative that the data be acquired only for certain duration on the occurrence of the event. Such acquisition systems are classified as event driven data acquisition system and are widely used in fusion engineering community.Development and implementation of high speed event driven data acquisition systems pose challenges in terms of data buffering, data streaming and synchronization. The precise timing of the event occurrence with respect to the main pulse should also be known for which a timing and synchronization system is needed. To fulfill this purpose, IEEE 1588 protocol offers a network based method of synchronizing multiple nodes based on Ethernet with sub microsecond accuracy. The present work describes the development of an event driven high speed data acquisition in synchronization with IEEE 1588 based timing network. The work proposes the time stamping of the event from synchronized network integrated with the event acquisition platform. The developed system has been tested for pulse durations longer than 3600 s. The performance of event data acquisition system has been thoroughly characterized and improved by applying various software optimizations in LabVIEW. The present work although intended to be used for INTF can be used for any such facility in fusion or other areas.The paper provides results of IEEE 1588 based timing network in various configurations and event data acquisition system developed on high speed FPGA controlled digitizer board with prototype results.

FastHGNN: A New Sampling Technique for Learning with Hypergraph Neural Networks

Hypergraphs can represent higher-order relations among objects. Traditional hypergraph neural networks involve node-edge-node transform, leading to high computational cost and timing. The main aim of this paper is to propose a new sampling technique for learning with hypergraph neural networks. The core idea is to design a layer-wise sampling scheme for nodes and hyperedges to approximate the original hypergraph convolution. We rewrite hypergraph convolution in the form of double integral and leverage Monte Carlo to achieve a discrete and consistent estimator. In addition, we use importance sampling and finally derive feasible probability mass functions for both nodes and hyperedges in consideration of variance reduction, based on some assumptions. Notably, the proposed sampling technique allows us to handle large-scale hypergraph learning, which is not feasible with traditional hypergraph neural networks. Experiment results demonstrate that our proposed model keeps a good balance between running time and prediction accuracy.