An Empirical Configuration Study of a Common Document Clustering Pipeline by Anton Eklund, Mona Forsman, Frank Drewes

“…The results suggest that BERT embeddings combined with UMAP dimension reduction to no less than 15 dimensions provides a good basis for clustering, regardless of the specific clustering algorithm used. Moreover, while UMAP performed better than PCA in our experiments, tuning the UMAP settings showed little impact on the overall performance. …”
Get full text

Accelerating and Compressing Transformer-Based PLMs for Enhanced Comprehension of Computer Terminology by Jian Peng, Kai Zhong

“…Our method involves a pipeline parallelism algorithm designed to accelerate training. It is paired with an innovative mixed compression strategy that combines pruning and knowledge distillation to effectively reduce the model size while preserving its performance. …”
Get full text

Secondary Intraocular Lens Options in Intraocular Lens Dislocation: A Review by Alexander Maloof, Ava James, Nicholas Pavic, Shane Zhang, Stephanie Wu, Christopher Go

“…While there are numerous IOL techniques and lens options for addressing IOL dislocation, there is currently no consensus on the optimal method. …”
Get full text

Design of Asymmetrical Relay Resonators for Maximum Efficiency of Wireless Power Transfer by Bo-Hee Choi, Jeong-Hae Lee

“…The optimum placement is found by scanning the positions of the relays and optimum capacitance can be found by using genetic algorithm (GA). The PTEs are enhanced when capacitance is optimally designed by GA according to the position of relays, respectively, and then maximum efficiency is obtained at the optimum placement of relays. …”
Get full text

Design of nodes importance assessment method for complex network based on neighborhood information by Xing LI, Jie ZHAN, Baoquan REN, Siqi ZHU

“…Accurate identification of influential nodes in complex networks is crucial for network management and network security.The local centrality method is concise and easy to use, but ignores the topological relationship between neighboring nodes and cannot provide globally optimal results.A node importance assessment method was proposed to correlate the node edge relationship and topology, which firstly applied the H-index and information entropy to assess the nodes, then combined the structural holes of the nodes with the structural characteristics of the nodes, and took into account the attribute of “bridging” while focusing on the node’s own quality and the amount of information about the neighboring nodes.The algorithm was validated by simulating the propagation process using the SIR model, and the Kendall correlation coefficient, complementary cumulative distribution function and propagation influence were applied to validate the validity and applicability of the method.Comparison of the experimental results on six real network datasets shows that the proposed method is more accurate than the traditional centrality methods in identifying and ordering the key nodes in the network.…”
Get full text

Adaptive rumor propagation and activity contagion in higher-order networks by Yafang Dong, Liang’an Huo, Matjaž Perc, Stefano Boccaletti

“…We also apply a sequential quadratic programming algorithm to optimize the parameters of our model and validate its accuracy and applicability on real-world data. …”
Get full text

Skip-Connected CNN Exploiting BNN Surrogate for Antenna Modelling by Yubo Tian, Jinlong Sun, Zhiwei Zhu

“…The combination of evolutionary algorithms with full-wave electromagnetic (EM) simulation software is an effective method for antenna optimal design. …”
Get full text

Client grouping and time-sharing scheduling for asynchronous federated learning in heterogeneous edge computing environment by Qianpiao MA, Qingmin JIA, Jianchun LIU, Hongli XU, Renchao XIE, Tao HUANG

“…., edge heterogeneity, data Non-IID, and communication resource constraints, a grouping asynchronous federated learning (FedGA) mechanism was proposed.Edge nodes were divided into multiple groups, each of which performed global updated asynchronously with the global model, while edge nodes within a group communicate with the parameter server through time-sharing communication.Theoretical analysis established a quantitative relationship between the convergence bound of FedGA and the data distribution among the groups.A time-sharing scheduling magic mirror method (MMM) was proposed to optimize the completion time of a single round of model updating within a group.Based on both the theoretical analysis for FedGA and MMM, an effective grouping algorithm was designed for minimizing the overall training completion time.Experimental results demonstrate that the proposed FedGA and MMM can reduce model training time by 30.1%～87.4% compared to the existing state-of-the-art methods.…”
Get full text

A hybrid unsupervised machine learning model with spectral clustering and semi-supervised support vector machine for credit risk assessment. by Tao Yu, Wei Huang, Xin Tang, Duosi Zheng

“…The effectiveness, efficiency, and robustness of the proposed TSC-SVM model are demonstrated through various real-world applications. The proposed algorithm is anticipated to expand the customer base for financial institutions while reducing economic losses.…”
Get full text

A framework for detecting and predicting highway traffic anomalies via multimodal fusion and heterogeneous graph neural networks. by Shaowei Sun, Mingzhou Liu

“…To address these limitations, we propose a framework based on multimodal deep fusion and heterogeneous graph neural networks (HGNNs), incorporating an Ensemble Contrastive Pessimistic Likelihood Estimation (CPLE) algorithm to optimize performance. The framework integrates static and dynamic traffic data, such as video images, traffic flow, vehicle speed, and tunnel weather conditions. …”
Get full text

Computing mutual similarity of 3D human faces in nearly linear time. by Radek Ošlejšek, Petra Urbanová, Jiří Sochor

“…Our solution utilizes properties of facial scan geometry to optimize individual steps. Moreover, the algorithm deals with possible holes and other artifacts in polygonal meshes automatically. …”
Get full text

Benthos-DETR: a high-precision efficient network for benthic organisms detection by Weibo Rao, Gang Chen, Yifei Zhang, Jue Cang, Shusen Chen, Chenyang Wang

“…Through a series of experimental analyses, it was evident that the performance of the Benthos-DETR network surpasses existing target detection algorithms, achieving an optimal equilibrium between high recognition precision and a trim network scale.…”
Get full text

Integrated multiomics analysis and machine learning refine neutrophil extracellular trap-related molecular subtypes and prognostic models for acute myeloid leukemia by Fangmin Zhong, Fangyi Yao, Zihao Wang, Jing Liu, Bo Huang, Xiaozhong Wang

“…The formation of NETs is impaired in acute myeloid leukemia (AML), which can result in immunodeficiency and increased susceptibility to infection.MethodsThe gene set variation analysis (GSVA) algorithm was employed for the calculation of NET score, while the consensus clustering algorithm was utilized to identify molecular subtypes. …”
Get full text

Minimization of Resource Consumption with URLLC Constraints for Relay-Assisted IIoT by Yujie Zhao, Tao Peng, Yichen Guo, Yijing Niu, Wenbo Wang

“…In this work, the packet error probability (PEP) and blocklength allocation across two-hop links are jointly optimized to minimize total blocklength (resource consumption) while satisfying reliability, latency, and throughput requirements. …”
Get full text

A Model‐Independent Strategy for the Targeted Observation Analysis and Its Application in ENSO Prediction by Weixun Rao, Youmin Tang, Yanling Wu, Xiaojing Li

“…Sensitive experiments show that while number of data sets reaches 12, a robust optimal observational array is obtained. …”
Get full text

Deep reinforcement learning applications and prospects in industrial scenarios by JING TAN, Ligang YANG, Xiaorui LI, Zhaolin YUAN, Yunduan CUI, Chao YAO, Zongjie WANG, Xiaojuan BAN

“…Central to these systems are control algorithms, which enable the automation of operations, optimization of process parameters, and reduction of operational costs. …”
Get full text

A Blockchain-Based Architecture of Web3.0: A Comprehensive Decentralized Model With Relay Nodes, Unique IDs and P2P by Hyunjoo Yang, Sejin Park

“…When faced with malicious requests, the system significantly reduces verification latency to an average of 0.0106 ms, ensuring rapid rejection and preserving network stability, while legitimate requests are processed efficiently with an average latency of 0.4125 ms.Experiments comparing latency-based and random-based relay selection methods across various network configurations (public-public, public-private, private-private) demonstrate that latency-based selection optimizes connection setup times in low-latency environments, while random-based selection provides robust performance in complex network scenarios. …”
Get full text

Adaptive Switching Redundant-Mode Multi-Core System for Photovoltaic Power Generation by Liang Liu, Xige Zhang, Jiahui Zhou, Kai Niu, Zixuan Guo, Yawen Zhao, Meng Zhang

“…Additionally, by analyzing the relationship between performance and reliability, we proposed optimization methods to improve reliability while ensuring a high performance was maintained. …”
Get full text

Novel channel attention-based filter pruning methods for low-complexity semantic segmentation models by Md. Bipul Hossain, Na Gong, Mohamed Shaban

“…Examples of widely used semantic segmentation models are the U-Net and DeeplabV3+ models. While the aforementioned models have been deemed very successful in segmenting medical targets including organs and diseases in high resolution images, the computational complexity represents a burden for the real-time application of the algorithms or the deployment of the models on resource-constrained platforms. …”
Get full text

Towards a smart debt collection system: a Design Science Research approach by Michał Przybyłek, Adam Przybyłek, Konrad Wojciechowski, Illia Shkroba

“…Abstract While manual, expert-based approaches to debt collection have long dominated the industry, the growing complexity and volume of data—resulting in information overload—highlight the need for machine-learning solutions. …”
Get full text