Hongli Xu
YOU?
Author Swipe
View article: Enabling Reconfiguration-Communication Overlap for Collective Communication in Optical Networks
Enabling Reconfiguration-Communication Overlap for Collective Communication in Optical Networks Open
Collective communication (CC) is widely adopted for large-scale distributed machine learning (DML) training workloads. DML's predictable traffic pattern provides a great oppotunity for applying optical network technology. Existing optical …
View article: Towards Communication-Efficient Decentralized Federated Graph Learning over Non-IID Data
Towards Communication-Efficient Decentralized Federated Graph Learning over Non-IID Data Open
Decentralized Federated Graph Learning (DFGL) overcomes potential bottlenecks of the parameter server in FGL by establishing a peer-to-peer (P2P) communication network among workers. However, while extensive cross-worker communication of g…
View article: Characteristics and Risk Factors of Inadvertent Intraoperative Hypothermia in Pediatric Surgery: A Study in Zhejiang Province, China
Characteristics and Risk Factors of Inadvertent Intraoperative Hypothermia in Pediatric Surgery: A Study in Zhejiang Province, China Open
Background: This study aims to explore the characteristics and risk factors of inadvertent intraoperative hypothermia (IIH) in pediatric surgical patients. Methods: Data from 291 pediatric surgical patients were retrospectively analyzed. D…
View article: Chain reaction mechanism study of active groups in coal
Chain reaction mechanism study of active groups in coal Open
View article: Mitigating Catastrophic Forgetting with Adaptive Transformer Block Expansion in Federated Fine-Tuning
Mitigating Catastrophic Forgetting with Adaptive Transformer Block Expansion in Federated Fine-Tuning Open
Federated fine-tuning (FedFT) of large language models (LLMs) has emerged as a promising solution for adapting models to distributed data environments while ensuring data privacy. Existing FedFT methods predominantly utilize parameter-effi…
View article: Cross-region Model Training with Communication-Computation Overlapping and Delay Compensation
Cross-region Model Training with Communication-Computation Overlapping and Delay Compensation Open
Training large language models (LLMs) requires massive computational resources, often necessitating the aggregation of geographically distributed data centers (\ie, cross-region training). However, the high communication latency in wide-ar…
View article: X-ClusterLink: An Efficient Cross-Cluster Communication Framework in Multi-Kubernetes Clusters
X-ClusterLink: An Efficient Cross-Cluster Communication Framework in Multi-Kubernetes Clusters Open
View article: Resource-Efficient Federated Fine-Tuning Large Language Models for Heterogeneous Data
Resource-Efficient Federated Fine-Tuning Large Language Models for Heterogeneous Data Open
Fine-tuning large language models (LLMs) via federated learning, i.e., FedLLM, has been proposed to adapt LLMs for various downstream applications in a privacy-preserving way. To reduce the fine-tuning costs on resource-constrained devices…
View article: Calibration and Verification of Coated Caragana korshinskii Seeds Based on Discrete Element Method
Calibration and Verification of Coated Caragana korshinskii Seeds Based on Discrete Element Method Open
The accurate modeling of seed motion characteristics is crucial for optimizing seed-metering devices in agricultural machinery. This study investigated the physical properties and contact parameters of coated Caragana korshinskii seeds thr…
View article: Lightweight and Post-Training Structured Pruning for On-Device Large Lanaguage Models
Lightweight and Post-Training Structured Pruning for On-Device Large Lanaguage Models Open
Considering the hardware-friendly characteristics and broad applicability, structured pruning has emerged as an efficient solution to reduce the resource demands of large language models (LLMs) on resource-constrained devices. Traditional …
View article: Efficient Deployment of Large Language Models on Resource-constrained Devices
Efficient Deployment of Large Language Models on Resource-constrained Devices Open
Deploying Large Language Models (LLMs) on resource-constrained (or weak) devices presents significant challenges due to limited resources and heterogeneous data distribution. To address the data concern, it is necessary to fine-tune LLMs u…
View article: Research on Adaptive Control Strategy for Co2 Heat Pump Air Conditioning System of Electric Vehicles Based on Artificial Neural Network and Genetic Algorithm
Research on Adaptive Control Strategy for Co2 Heat Pump Air Conditioning System of Electric Vehicles Based on Artificial Neural Network and Genetic Algorithm Open
View article: A Robust Federated Learning Framework for Undependable Devices at Scale
A Robust Federated Learning Framework for Undependable Devices at Scale Open
In a federated learning (FL) system, many devices, such as smartphones, are often undependable (e.g., frequently disconnected from WiFi) during training. Existing FL frameworks always assume a dependable environment and exclude undependabl…
View article: Caesar: A Low-deviation Compression Approach for Efficient Federated Learning
Caesar: A Low-deviation Compression Approach for Efficient Federated Learning Open
Compression is an efficient way to relieve the tremendous communication overhead of federated learning (FL) systems. However, for the existing works, the information loss under compression will lead to unexpected model/gradient deviation f…
View article: Adaptive Parameter-Efficient Federated Fine-Tuning on Heterogeneous Devices
Adaptive Parameter-Efficient Federated Fine-Tuning on Heterogeneous Devices Open
Federated fine-tuning (FedFT) has been proposed to fine-tune the pre-trained language models in a distributed manner. However, there are two critical challenges for efficient FedFT in practical applications, i.e., resource constraints and …
View article: Enhancing Federated Graph Learning via Adaptive Fusion of Structural and Node Characteristics
Enhancing Federated Graph Learning via Adaptive Fusion of Structural and Node Characteristics Open
Federated Graph Learning (FGL) has demonstrated the advantage of training a global Graph Neural Network (GNN) model across distributed clients using their local graph data. Unlike Euclidean data (\eg, images), graph data is composed of nod…
View article: Collaborative Inference for Large Models with Task Offloading and Early Exiting
Collaborative Inference for Large Models with Task Offloading and Early Exiting Open
In 5G smart cities, edge computing is employed to provide nearby computing services for end devices, and the large-scale models (e.g., GPT and LLaMA) can be deployed at the network edge to boost the service quality. However, due to the con…
View article: Many Hands Make Light Work: Accelerating Edge Inference via Multi-Client Collaborative Caching
Many Hands Make Light Work: Accelerating Edge Inference via Multi-Client Collaborative Caching Open
Edge inference is a technology that enables real-time data processing and analysis on clients near the data source. To ensure compliance with the Service-Level Objectives (SLOs), such as a 30% latency reduction target, caching is usually a…
View article: SRSA: A Cost-Efficient Strategy-Router Search Agent for Real-world Human-Machine Interactions
SRSA: A Cost-Efficient Strategy-Router Search Agent for Real-world Human-Machine Interactions Open
Recently, as Large Language Models (LLMs) have shown impressive emerging capabilities and gained widespread popularity, research on LLM-based search agents has proliferated. In real-world situations, users often input contextual and highly…
View article: Anti-NMDAR encephalitis with delayed ovarian teratoma in a young woman: a case report with 5 years of follow-up
Anti-NMDAR encephalitis with delayed ovarian teratoma in a young woman: a case report with 5 years of follow-up Open
View article: ParallelSFL: A Novel Split Federated Learning Framework Tackling Heterogeneity Issues
ParallelSFL: A Novel Split Federated Learning Framework Tackling Heterogeneity Issues Open
Mobile devices contribute more than half of the world's web traffic, providing massive and diverse data for powering various federated learning (FL) applications. In order to avoid the communication bottleneck on the parameter server (PS) …
View article: Research on the Integration of Curriculum Ideology and Politics into the Teaching Process of Mechanical Manufacturing Engineering
Research on the Integration of Curriculum Ideology and Politics into the Teaching Process of Mechanical Manufacturing Engineering Open
Curriculum ideology and politics is a new model of college professional curriculum education, that is, by integrating the elements of thought and politics into the form of professional courses to effectively guide students to establish a c…
View article: Automatic Crack Detection Using Weakly Supervised Semantic Segmentation Network and Mixed-Label Training Strategy
Automatic Crack Detection Using Weakly Supervised Semantic Segmentation Network and Mixed-Label Training Strategy Open
Automatic crack detection in construction facilities is a challenging yet crucial task. However, existing deep learning (DL)-based semantic segmentation methods for this field are based on fully supervised learning models and pixel-level m…
View article: Key to the species of the genus Subancistrocerus de Saussure, 1855 (Hymenoptera, Vespidae, Eumeninae) from China with description of a new species
Key to the species of the genus Subancistrocerus de Saussure, 1855 (Hymenoptera, Vespidae, Eumeninae) from China with description of a new species Open
A newly discovered species, Subancistrocerus clypeatus sp. nov. , from China (Zhejiang) is described and illustrated. In addition, Subancistrocerus kankauensis (Schulthess-Rechberg) is redescribed and photographed after studying the type s…
View article: Locally Differentially Private Heterogeneous Graph Aggregation with Utility Optimization
Locally Differentially Private Heterogeneous Graph Aggregation with Utility Optimization Open
Graph data are widely collected and exploited by organizations, providing convenient services from policy formation and market decisions to medical care and social interactions. Yet, recent exposures of private data abuses have caused huge…
View article: Key to the species of the genus Subancistrocerus de Saussure, 1855 (Hymenoptera, Vespidae, Eumeninae) from China with description of a new species
Key to the species of the genus Subancistrocerus de Saussure, 1855 (Hymenoptera, Vespidae, Eumeninae) from China with description of a new species Open
This dataset contains the digitized treatments in Plazi based on the original journal article Tan, Jiang-Li, Wang, Meng, Xu, Hongli, Tang, Yan, Liu, Ying (2023): Key to the species of the genus Subancistrocerus de Saussure, 1855 (Hymenopte…
View article: Adaptive Control of Client Selection and Gradient Compression for Efficient Federated Learning
Adaptive Control of Client Selection and Gradient Compression for Efficient Federated Learning Open
Federated learning (FL) allows multiple clients cooperatively train models without disclosing local data. However, the existing works fail to address all these practical concerns in FL: limited communication resources, dynamic network cond…
View article: Adaptive Configuration for Heterogeneous Participants in Decentralized Federated Learning
Adaptive Configuration for Heterogeneous Participants in Decentralized Federated Learning Open
Data generated at the network edge can be processed locally by leveraging the paradigm of edge computing (EC). Aided by EC, decentralized federated learning (DFL), which overcomes the single-point-of-failure problem in the parameter server…
View article: Multimodal Teaching of Japanese Newspaper Reading
Multimodal Teaching of Japanese Newspaper Reading Open
With the deepening of teaching reform and the development of network technology, single teaching mode has been difficult to adapt to the needs of modern teaching.As a kind of teaching theory, multimodal teaching believes that students shou…
View article: Research on the Use of Echoing Performance on Japanese TV Talk Shows
Research on the Use of Echoing Performance on Japanese TV Talk Shows Open