Convergence analysis of reweighted sum-product algorithms

被引:20
|
作者
Roosta, Tanya G. [1 ]
Wainwright, Martin J. [1 ,2 ]
Sastry, Shankar S. [1 ]
机构
[1] Univ Calif Berkeley, Dept Elect & Comp Engn, Berkeley, CA 94720 USA
[2] Univ Calif Berkeley, Dept Stat, Berkeley, CA 94720 USA
关键词
approximate marginalization; belief propagation; convergence analysis; graphical models; Markov random fields; sum-product algorithm;
D O I
10.1109/TSP.2008.924136
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Markov random fields are designed to represent structured dependencies among large collections of random variables, and are well-suited to capture the structure of real-world signals. Many fundamental tasks in signal processing (e.g., smoothing, denoising, segmentation etc.) require efficient methods for computing (approximate) marginal probabilities over subsets of nodes in the graph. The marginalization problem, though solvable in linear time for graphs without cycles, is computationally intractable for general graphs with cycles. This intractability motivates the use of approximate "message-passing" algorithms. This paper studies the convergence and stability properties of the family of reweighted sum-product algorithms, a generalization of the widely used sum-product or belief propagation algorithm, in which messages are adjusted with graph-dependent weights. For pairwise Markov random fields, we derive various conditions that are sufficient to ensure convergence, and also provide bounds on the geometric convergence rates. When specialized to the ordinary sum-product algorithm, these results provide strengthening of previous analyses. We prove that some of our conditions are necessary and sufficient for subclasses of homogeneous models, but not for general models. The experimental simulations on various classes of graphs validate our theoretical results.
引用
收藏
页码:4293 / 4305
页数:13
相关论文
共 50 条
  • [21] Classification-Aided Multitarget Tracking Using the Sum-Product Algorithm
    Gaglione, Domenico
    Soldi, Giovanni
    Braca, Paolo
    De Magistris, Giovanni
    Meyer, Florian
    Hlawatsch, Franz
    IEEE SIGNAL PROCESSING LETTERS, 2020, 27 : 1710 - 1714
  • [22] Efficient Maximum Likelihood Estimation for Pedigree Data with the Sum-Product Algorithm
    Engelhardt, Alexander
    Rieger, Anna
    Tresch, Achim
    Mansmann, Ulrich
    HUMAN HEREDITY, 2016, 82 (1-2) : 1 - 15
  • [23] Boost Sum-Product Performance for Multiuser Detection in mMTC at Millimeter Wave
    Huang, Tao
    Ye, Baoliu
    Tang, Bin
    Xie, Lei
    Lu, Sanglu
    Guo, Song
    IEEE TRANSACTIONS ON MOBILE COMPUTING, 2023, 22 (02) : 765 - 780
  • [24] HETEROGENEOUS INFORMATION FUSION FOR MULTITARGET TRACKING USING THE SUM-PRODUCT ALGORITHM
    Soldi, Giovanni
    Gaglione, Domenico
    Meyer, Florian
    Hlawatsch, Franz
    Braca, Paolo
    Farina, Alfonso
    Win, Moe Z.
    2019 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2019, : 5471 - 5475
  • [25] Distribution of Statistics of Hidden State Sequences Through the Sum-Product Algorithm
    Donald E. K. Martin
    John A. D. Aston
    Methodology and Computing in Applied Probability, 2013, 15 : 897 - 918
  • [26] 3-D neural mapper for LDPC sum-product decoder
    Boiko Y.
    Optical Memory and Neural Networks (Information Optics), 2009, 18 (02): : 101 - 107
  • [27] Performance evaluation of a modified sum-product decoding algorithm for LDPC codes
    Papaharalabos, S
    Sweeney, P
    Evans, BG
    Albertazzi, G
    Vanelli-Coralli, A
    Corazza, GE
    2nd International Symposium on Wireless Communications Systems 2005 (ISWCS 2005), 2005, : 800 - 804
  • [28] Distribution of Statistics of Hidden State Sequences Through the Sum-Product Algorithm
    Martin, Donald E. K.
    Aston, John A. D.
    METHODOLOGY AND COMPUTING IN APPLIED PROBABILITY, 2013, 15 (04) : 897 - 918
  • [29] A novel broadcast scheduling strategy using factor graphs and the sum-product algorithm
    Chen, Jung-Chieh
    Wang, Yeong-Cheng
    Chen, Jiunn-Tsair
    IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, 2006, 5 (06) : 1241 - 1249
  • [30] Belief Propagation Estimation of Protein and Domain Interactions Using the Sum-Product Algorithm
    Morcos, Faruck
    Sikora, Marcin
    Alber, Mark S.
    Kaiser, Dale
    Izaguirre, Jesus A.
    IEEE TRANSACTIONS ON INFORMATION THEORY, 2010, 56 (02) : 742 - 755