Explanipedia

FedQuad: Adaptive Layer-wise LoRA Deployment and Activation Quantization for Federated Fine-Tuning Open

Rukuo Li, Jianchun Liu, Hongli Xu, Liusheng Huang · 2025

Federated fine-tuning (FedFT) provides an effective paradigm for fine-tuning large language models (LLMs) in privacy-sensitive scenarios. However, practical deployment remains challenging due to the limited resources on end devices. Existi…

Cross-region Model Training with Communication-Computation Overlapping and Delay Compensation Open

Ying Zhu, Xu Yang, Hongli Xu, Yunming Liao, Zhiwei Yao , et al. · 2025

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…

Stop Diverse OOD Attacks: Knowledge Ensemble for Reliable Defense Open

Zhenbo Shi, Xiaoman Liu, Yuxuan Zhang, Shuchang Wang, Rui Shu , et al. · 2025

Enhancing defense through model ensemble is an emerging trend, where the challenge lies in how to use ensemble knowledge to counter Out-of-Distribution (OOD) attacks. In this paper, we propose the Reliable Defense Ensemble (REE) to address…

Collaborative Speculative Inference for Efficient LLM Inference Serving Open

Luyao Gao, Jianchun Liu, Yang Xu, X. Zhang, Yunming Liao , et al. · 2025

Speculative inference is a promising paradigm employing small speculative models (SSMs) as drafters to generate draft tokens, which are subsequently verified in parallel by the target large language model (LLM). This approach enhances the …

Top-n𝜎: Eliminating Noise in Logit Space for Robust Token Sampling of LLM Open

C. J. Tang, Jianchun Liu, Hongli Xu, Liusheng Huang · 2025

Enhancing Federated Graph Learning via Adaptive Fusion of Structural and Node Characteristics Open

Xianjun Gao, Jianchun Liu, Hongli Xu, Shilong Wang, Liusheng Huang · 2024

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…

Accelerating End-Cloud Collaborative Inference via Near Bubble-free Pipeline Optimization Open

Luyao Gao, Jianchun Liu, Hongli Xu, Xu Sun, Qianpiao Ma , et al. · 2024

End-cloud collaboration offers a promising strategy to enhance the Quality of Service (QoS) in DNN inference by offloading portions of the inference workload from end devices to cloud servers. Despite the potential, the complex model archi…

Many Hands Make Light Work: Accelerating Edge Inference via Multi-Client Collaborative Caching Open

Wanting Liang, Jianchun Liu, Hongli Xu, Chunming Qiao, Liusheng Huang · 2024

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…

ParallelSFL: A Novel Split Federated Learning Framework Tackling Heterogeneity Issues Open

Yunming Liao, Yang Xu, Hongli Xu, Zhiwei Yao, Liusheng Huang , et al. · 2024

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) …

Locally Differentially Private Heterogeneous Graph Aggregation with Utility Optimization Open

Zichun Liu, Liusheng Huang, Hongli Xu, Wei Yang · 2023

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…

Shape Prior Guided Attack: Sparser Perturbations on 3D Point Clouds Open

Zhenbo Shi, Zhi Chen, Zhenbo Xu, Wei Yang, Zhidong Yu , et al. · 2022

Deep neural networks are extremely vulnerable to malicious input data. As 3D data is increasingly used in vision tasks such as robots, autonomous driving and drones, the internal robustness of the classification models for 3D point cloud h…

F3SNet: A Four-Step Strategy for QIM Steganalysis of Compressed Speech Based on Hierarchical Attention Network Open

Chuanpeng Guo, Wei Yang, Mengxia Shuai, Liusheng Huang · 2021

Traditional machine learning-based steganalysis methods on compressed speech have achieved great success in the field of communication security. However, previous studies lacked mathematical modeling of the correlation between codewords, a…

F3SNet: A Four-Step Strategy for QIM Steganalysis of Compressed Speech Based on Hierarchical Attention Network Open

Chuanpeng Guo, Wei Yang, Liusheng Huang · 2021

Traditional machine learning-based steganalysis methods on compressed speech have achieved great success in the field of communication security. However, previous studies lacked mathematical description and modeling of the correlation betw…

PointTrack++ for Effective Online Multi-Object Tracking and Segmentation Open

Zhenbo Xu, Wei Zhang, Xiao Tan, Wei Yang, Xiangbo Su , et al. · 2020

Multiple-object tracking and segmentation (MOTS) is a novel computer vision task that aims to jointly perform multiple object tracking (MOT) and instance segmentation. In this work, we present PointTrack++, an effective on-line framework f…

Segment as Points for Efficient Online Multi-Object Tracking and Segmentation Open

Zhenbo Xu, Wei Zhang, Xiao Tan, Wei Yang, Huan Huang , et al. · 2020

Current multi-object tracking and segmentation (MOTS) methods follow the tracking-by-detection paradigm and adopt convolutions for feature extraction. However, as affected by the inherent receptive field, convolution based feature extracti…

ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection Open

Zhenbo Xu, Wei Zhang, Xiaoqing Ye, Xiao Tan, Wei Yang , et al. · 2020

3D object detection is an essential task in autonomous driving and robotics. Though great progress has been made, challenges remain in estimating 3D pose for distant and occluded objects. In this paper, we present a novel framework named Z…

ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection Open

Zhenbo Xu, Wei Zhang, Xiaoqing Ye, Xiao Tan, Wei Yang , et al. · 2020

3D object detection is an essential task in autonomous driving and robotics. Though great progress has been made, challenges remain in estimating 3D pose for distant and occluded objects. In this paper, we present a novel framework named Z…

Aggregating Votes with Local Differential Privacy: Usefulness, Soundness vs. Indistinguishability Open

Shaowei Wang, Jiachun Du, Wei Yang, Xinrong Diao, Zichun Liu , et al. · 2019

Voting plays a central role in bringing crowd wisdom to collective decision making, meanwhile data privacy has been a common ethical/legal issue in eliciting preferences from individuals. This work studies the problem of aggregating indivi…

How Bob in Quantum Private Query Protocol Gets the Element? Open

Bingren Chen, Lide Xue, Wei Yang, Liusheng Huang · 2019

Quantum private query (QPQ) requires that the database holder Bob knows nothing about his client Alice, including the index she provides and the element she obtains. However, on some occasion, Bob wants to know which element he has reveale…

Attention-based Walking Gait and Direction Recognition in Wi-Fi Networks Open

Yang Xu, Min Chen, Wei Yang, Sheng Chen, Liusheng Huang · 2018

The study of human gait recognition has been becoming an active research field. In this paper, we propose to adopt the attention-based Recurrent Neural Network (RNN) encoder-decoder framework to implement a cycle-independent human gait and…

COUSTIC: Combinatorial Double auction for Task Assignment in Device-to-Device Clouds Open

Yutong Zhai, Liusheng Huang, Long Chen, Ning Xiao, Yangyang Geng · 2018

With the emerging technologies of Internet of Things (IOTs), the capabilities of mobile devices have increased tremendously. However, in the big data era, to complete tasks on one device is still challenging. As an emerging technology, cro…

Hybrid routing by joint optimization of per-flow routing and tag-based routing in software-defined networks Open

Gongming Zhao, Liusheng Huang, Ziqiang Li, Hongli Xu · 2018

In recent years, Software-Defined Networks (SDNs) have become a promising technology to improve network utilization. However, limited flow table size and long deployment delays may result in low network performance in large-scale networks …

Incorporating Latent Meanings of Morphological Compositions to Enhance Word Embeddings Open

Yang Xu, Jiawei Liu, Wei Yang, Liusheng Huang · 2018

Traditional word embedding approaches learn semantic information at word level while ignoring the meaningful internal structures of words like morphemes. Furthermore, existing morphology-based models directly incorporate morphemes to train…

Cost-Effective Seed Selection in Online Social Networks Open

Kai Han, Yuntian He, Xiaokui Xiao, Shaojie Tang, Jingxin Xu , et al. · 2017

We study the min-cost seed selection problem in online social networks, where the goal is to select a set of seed nodes with the minimum total cost such that the expected number of influenced nodes in the network exceeds a predefined thres…

Personalized Classifier Ensemble Pruning Framework for Mobile Crowdsourcing Open

Shaowei Wang, Liusheng Huang, Pengzhan Wang, Hongli Xu, Wei Yang · 2017

Ensemble learning has been widely employed by mobile applications, ranging from environmental sensing to activity recognitions. One of the fundamental issue in ensemble learning is the trade-off between classification accuracy and computat…

Optimizing virtual machine placement in distributed clouds with M/M/1 servers Open

Hou Deng, Liusheng Huang, Chenkai Yang, Hongli Xu, Bing Leng · 2017

Personalized Differential Privacy Preserving Data Aggregation for Smart Homes Open

Xinyuan Zhang, Liusheng Huang, Shaowei Wang, Zhenyu Zhu, Hongli Xu · 2017

The aggregation of residents' private data drives improvements in the smart homes, however it comes with compromising on privacy.Hence, privacy preservation has become an increasing requirement for residents.Since users might have differen…

A Hybrid Covert Channel Over LTE-A System Open

Zukui Wang, Liusheng Huang, Wei Yang, Zhiqiang He · 2017

In this paper, we have analyzed the sub-protocol stack of Long Term Evolution Advanced (LTE-A) System, and proposed a hybrid covert channel, called HyLTEsteg, designed for LTE-A System.The HyLTEsteg uses covert timing channels (CTC), which…

DYPSOKM: A Dynamic Union Of PSO And K-Means, A Better Cluster Open

Ke Qin, Liusheng Huang, Hongli Xu · 2017

As one of the most famous clustering algorithms, K-means is simple and effective but easily falls into local optimal solution.Aimed at this flaw, many methods including PSO had been applied to optimize K-means.As a typical swarm intelligen…

Recognizing the Operating Hand and the Hand-Changing Process for User Interface Adjustment on Smartphones Open

Hansong Guo, He Huang, Liusheng Huang, Yu-E Sun · 2016

As the size of smartphone touchscreens has become larger and larger in recent years, operability with a single hand is getting worse, especially for female users. We envision that user experience can be significantly improved if smartphone…

Liusheng Huang YOU? Author Swipe