Article
Engineering, Electrical & Electronic
Man Luo, Qinghua Guo, Ming Jin, Yonina C. Eldar, Defeng Huang, Xiangming Meng
Summary: In this work, a new SBL algorithm based on structured variational inference leveraging AMP with a unitary transformation is proposed. The proposed UAMP-SBL is shown to be more robust and efficient compared to state-of-the-art AMP-based SBL algorithms, leading to remarkably better performance.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Qiuyun Zou, Haochuan Zhang, Hongwen Yang
Summary: This paper extends the BiG-AMP method to handle matrix factorization problems and introduces the ML-BiGAMP algorithm, which can achieve the same performance as traditional methods but with lower computational burden. The performance of ML-BiGAMP is fully characterized through a set of one-dimensional equations termed state evolution (SE).
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Linlin Mo, Xinhua Lu, Jide Yuan, Chuanzong Zhang, Zhongyong Wang, Petar Popovski
Summary: The double linear transformation model Y = AXB + W plays an important role in various science and engineering applications. In this work, a generalized algorithm called GD-UAMP is developed to solve the decoupling problem between X and Y, by incorporating double unitary transformation. Numerical results show that GD-UAMP outperforms benchmarks in channel estimation for wireless communication systems.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Hui Zhang, Shoujiang Li, Yong Liang, Hai Zhang, Mengmeng Du
Summary: This paper investigates efficient algorithms for nonconvex regularization methods using deep learning, and proposes neural network architectures based on AMP and VAMP algorithms. The effectiveness of these networks is demonstrated through experiments.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Natalia M. Arzeno, Haris Vikalo
Summary: Evolutionary affinity propagation (EAP) is introduced as an evolutionary clustering algorithm that automatically determines the number of clusters and tracks them, providing effective solutions for clustering time-evolving data. The proposed EAP algorithm demonstrates effectiveness through comparison with existing methods on simulated and experimental data.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2021)
Article
Engineering, Electrical & Electronic
Dan Zhang, Xiaohang Song, Wenjin Wang, Gerhard Fettweis, Xiqi Gao
Summary: This article unifies variational message passing, belief propagation, and expectation propagation under an optimization framework of Bethe free energy minimization with differently imposed constraints, providing a theoretical framework for systematically deriving message passing variants. By reformulating constraints, a low-complexity EP variant is obtained for better estimation performance. Furthermore, a hybrid message passing algorithm is systematically derived for joint SSR and statistical model learning with near-optimal inference performance and scalable complexity.
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2021)
Article
Engineering, Electrical & Electronic
Suwen Song, Zhongfeng Wang
Summary: This paper introduces a massive MIMO detection scheme based on message passing detection (MPD) and proposes an improved layered MPD (ILMPD) algorithm. By introducing a layered updating schedule and algorithmic transformations or approximations, the complexity is reduced. Moreover, a lightweight early termination strategy is explored to reduce the number of detection iterations. Based on the proposed algorithm, an area-efficient architecture is designed, and a reconfigurable version of the detector is developed.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
(2022)
Article
Engineering, Electrical & Electronic
Jingyi Dong, Wentao Lyu, Di Zhou, Weiqiang Xu
Summary: In this paper, a novel sparse Bayesian learning framework for large-scale image recovery is proposed. The VGAMP-SBL model, which combines variational Bayesian and generalized approximate message passing, is introduced to accelerate image reconstruction. Experimental results validate the effectiveness of the proposed method.
IEEE SIGNAL PROCESSING LETTERS
(2022)
Article
Engineering, Electrical & Electronic
Pei-Duo Yu, Chee Wei Tan, Hung-Lin Fu
Summary: This paper studies the epidemic source detection problem in contact tracing networks using the susceptible-infected model in epidemiology. By analyzing finite degree regular graphs and regular graphs with cycles, the mathematical equivalence between acyclic and cyclic graphs is established. A novel statistical distance centrality is proposed to refine the solution of the maximum likelihood estimator. The performance evaluation shows that the proposed algorithm outperforms existing heuristics in correctly identifying superspreaders in real infection clusters.
IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING
(2022)
Article
Computer Science, Information Systems
Jing Li, Tianming Yu, Ye Wang, Roger Wattenhofer
Summary: This paper studies the problem of dynamic reconfiguration in fully asynchronous message-passing systems with Byzantine faults. By proposing a dynamic Byzantine consistent broadcast algorithm and a dynamic Byzantine reliable broadcast algorithm, the solvability of dynamic Byzantine broadcast is demonstrated.
Article
Engineering, Electrical & Electronic
Feiyan Tian, Lei Liu, Xiaoming Chen
Summary: In this paper, a low-complexity and widely applicable generalized memory AMP (GMAMP) framework is proposed for signal reconstruction. The framework has high performance and applicability, and can be used to build new advanced AMP-type algorithms. A Bayes-optimal GMAMP (BO-GMAMP) algorithm is also introduced, which suppresses linear interference through a memory match filter estimator, with comparable complexity to existing algorithms.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2022)
Article
Engineering, Electrical & Electronic
Yabo Guo, Peng Sun, Zhengdao Yuan, Chongwen Huang, Qinghua Guo, Zhongyong Wang, Chau Yuen
Summary: Reconfigurable intelligent surface (RIS) has great potential in wireless networks for achieving high energy efficiency, extended coverage, improved capacity, and massive connectivity. However, accurately acquiring channel state information for RIS-aided communications is challenging. Existing channel estimation methods for RIS-aided multiple-input and multiple-output (MIMO) communications have high computational complexity and special requirements on matrices involved, hindering their applications. In this work, a new signal model is derived and a more efficient channel estimator based on unitary approximate message passing (UAMP) is developed, which has linear complexity with the number of RIS units and does not have special requirements on matrices. Extensive numerical results show that the proposed estimator outperforms existing methods and reduces training overhead and latency.
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2023)
Article
Engineering, Electrical & Electronic
Zhaorong Zhang, Minyue Fu
Summary: This paper studies the convergence properties of the message-passing algorithm for convex optimization, providing a simple bound for the convergence rate under the assumptions of pairwise separability and scaled diagonal dominance. In comparison with previous results, the paper does not require the convex program to have known convex pairwise components and offers a tighter and simpler bound for the convergence rate. Additionally, when specialized to quadratic optimization, the paper generalizes known results by providing a very simple bound for the convergence rate.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Zhenyu Zhang, Yuanyuan Dong, Keping Long, Xiyuan Wang, Xiaoming Dai
Summary: This paper proposes a decentralized Gaussian message passing detection method for decentralized baseband processing architecture. By iteratively calculating local means and variances and fusing them to generate global symbol beliefs, this method alleviates the issues of excessively high interconnect bandwidth and detection complexity in massive multi-user multiple-input multiple-output systems. Numerical analysis and simulation results verify the effectiveness and advantages of this method.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2022)
Article
Engineering, Electrical & Electronic
Alireza Sheikh, Alexandre Amat, Gianluigi Liva, Alex Alvarado
Summary: The proposed iBDD-CR algorithm enhances the conventional iBDD algorithm for product codes by utilizing soft information and reliability estimates. By analyzing the extrinsic message passing of GLDPC ensembles, which include product codes, iBDD-CR achieves performance gains up to 0.51 dB compared to iBDD. This makes it suitable for high-throughput applications like fiber-optic communications.
JOURNAL OF LIGHTWAVE TECHNOLOGY
(2021)