Article
Computer Science, Theory & Methods
Laurent Feuilloley, Pierre Fraigniaud
Summary: Proof-labeling schemes are mechanisms that provide nodes of networks with locally verifiable certificates. This paper introduces error-sensitive proof-labeling schemes that can detect network states far from satisfying given predicates with a linear proportion of detecting nodes.
JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING
(2022)
Article
Automation & Control Systems
Aixin Liu, Haitao Li
Summary: This paper introduces a new multi-potential equation based on the semi-tensor product of matrices for the verification of multi-potential games with a given partition, and presents new results on potential functions. An illustrative example demonstrates the effectiveness of the proposed approach.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2022)
Article
Automation & Control Systems
Shengchao He, Xiangdong Liu, Pingli Lu, Haikuo Liu, Changkun Du
Summary: This paper investigates the leader-follower finite-time consensus problem for multiagent systems with nonlinear dynamics using an intermittent protocol. The nonlinear dynamics of each agent satisfy Holder continuity, which is different from most existing works. By utilizing the finite-time control technique, an intermittent control protocol is designed to achieve accurate leader-follower finite-time consensus. The paper shows that the consensus can be realized if the length of communication exceeds a critical value through limit theory. Numerical examples are provided to validate the effectiveness of the proposed scheme.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Xiaoshuan Shi, Zhenhua Guo, Kang Li, Yun Liang, Xiaofeng Zhu
Summary: Noisy labels can significantly degrade the performance of convolutional neural networks (CNNs). This paper proposes a novel self-paced resistance framework to resist corrupted labels, using the memorization effect of CNNs and a resistance loss to update the model parameters. Extensive experiments demonstrate the superior performance of this framework on noisy-label data.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Theory & Methods
Bin Yu, Cong Tian, Xu Lu, Nan Zhang, Zhenhua Duan
Summary: This paper proposes a distributed network-based parallel runtime verification approach for checking the full regular temporal properties of C programs. Experimental results show that the proposed approach has a significant speedup compared to state-of-the-art approaches and supports more expressive properties.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2023)
Article
Computer Science, Theory & Methods
Danyang Xiao, Chengang Yang, Weigang Wu
Summary: Split Learning is a distributed machine learning setting that protects data privacy by allowing nodes to train neural networks based on model parallelism. However, recent studies show that raw data may be reconstructed from activations, posing a privacy risk. To address this, a mechanism called multiple activations and labels mix (MALM) is proposed, which generates mixed activations and obfuscated labels to reduce the risk of reconstruction attacks and prevent inference of ground-truth labels.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2022)
Article
Computer Science, Information Systems
Luis Miguel Danielsson, Cesar Sanchez
Summary: - We study the problem of monitoring distributed systems and propose a solution to monitor decentralized systems using stream runtime verification specifications.
- We describe the monitoring problem for timed-asynchronous networks and provide a decentralized algorithm along with proofs of its correctness.
Article
Computer Science, Artificial Intelligence
Fan Ma, Yu Wu, Xin Yu, Yi Yang
Summary: Deep neural networks are effective but data hungry in many applications. To address model performance degradation caused by erroneous labels, a novel reweighting method called SRCC is proposed, along with the use of mixed inputs to regularize decision boundaries. The method outperforms the state of the art on learning with noisy data.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Hardware & Architecture
Leonid Barenboim, Michael Elkin, Uri Goldenberg
Summary: This article discusses graph coloring and related problems in the distributed message-passing model. It introduces a locally-iterative Delta + 1-coloring algorithm that breaks the limitations of the Szegedy-Vishwanathan barrier, resulting in significant improvements for dynamic, self-stabilizing, and bandwidth-restricted settings.
JOURNAL OF THE ACM
(2022)
Article
Mathematics
Yu Yao, Guodong Zhang, Yan Li
Summary: This article explores complex-valued inertial neural networks with distributed delays and establishes novel results on fixed-time stabilization and preassigned-time stabilization by constructing two new feedback controllers. Unlike most previous works, the stabilization methods obtained here do not require dividing the original complex-valued system into separate real valued subsystems. Several numerical examples are provided to verify the effectiveness and reliability of the results, showing that the stabilization methods can be successfully implemented without being affected by system parameters and initial values.
Article
Computer Science, Artificial Intelligence
Chunxu Zhang, Ximing Li, Hongbin Pei, Zijian Zhang, Bing Liu, Bo Yang
Summary: In this study, a label-enhanced network architecture called LAENNet is proposed to address the robustness issue of GCNs in noisy and sparse graph data scenarios. Experimental results demonstrate the superiority of LAENNet over existing baseline models.
Article
Chemistry, Analytical
Meng Wang, Dazheng Feng, Tingting Su, Mohan Chen
Summary: This paper proposes a novel attention-based frequency aggregation method and two temporal-frequency aggregation methods for speaker verification systems. Experimental results show that CNN-based SV systems using these methods achieve significant improvements on Voxceleb dataset.
Article
Chemistry, Physical
Di Liu, Chengyu Li, Pengfei Chen, Xin Zhao, Wei Tang, Zhong Lin Wang
Summary: This study proposes a network system consisting of low-cost, maintenance-free, and distributed self-powered wireless monitoring nodes for long-term and wide-area environmental monitoring. The system is capable of automatically monitoring and wirelessly transmitting data in various natural areas.
ADVANCED ENERGY MATERIALS
(2023)
Article
Engineering, Biomedical
Olivier Petit, Nicolas Thome, Luc Soler
Summary: This paper proposes a method for handling partially labeled medical image datasets, which can effectively train neural networks with good robustness and performance. By introducing confidence self-training and multi-label loss strategies, it can effectively address the issue of missing pixel-level labels.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2021)
Article
Computer Science, Information Systems
Kun Li, Shengling Wang, Xiuzhen Cheng, Qin Hu
Summary: This paper proposes a misreport- and collusion-proof crowdsourcing mechanism that guides workers to truthfully report task quality, allowing quality control without verification. Simulation results verify the effectiveness of the proposed mechanism and reveal interesting findings that can impact strategic planning solutions for crowdsourcing.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2022)