首页> 外文OA文献 >Code optimization, frozen glassy phase and improved decoding algorithms for low-density parity-check codes
【2h】

Code optimization, frozen glassy phase and improved decoding algorithms for low-density parity-check codes

机译:代码优化,冷冻玻璃相和改进的解码算法   用于低密度奇偶校验码

摘要

The statistical physics properties of low-density parity-check codes for thebinary symmetric channel are investigated as a spin glass problem withmulti-spin interactions and quenched random fields by the cavity method. Byevaluating the entropy function at the Nishimori temperature, we find thatirregular constructions with heterogeneous degree distribution of check (bit)nodes have higher decoding thresholds compared to regular counterparts withhomogeneous degree distribution. We also show that the instability of themean-field calculation takes place only after the entropy crisis, suggestingthe presence of a frozen glassy phase at low temperatures. When no priorknowledge of channel noise is assumed (searching for the ground state), we findthat a reinforced strategy on normal belief propagation will boost the decodingthreshold to a higher value than the normal belief propagation. This value isclose to the dynamical transition where all local search heuristics fail toidentify the true message (codeword or the ferromagnetic state). After thedynamical transition, the number of metastable states with larger energydensity (than the ferromagnetic state) becomes exponentially numerous. When thenoise level of the transmission channel approaches the static transition point,there starts to exist exponentially numerous codewords sharing the identicalferromagnetic energy.
机译:通过腔法,研究了具有多旋相互作用和淬灭随机场的自旋玻璃问题,研究了双对称通道低密度奇偶校验码的统计物理性质。通过评估Nishimori温度下的熵函数,我们发现具有校验(位)节点异构度分布的不规则结构与具有均匀度分布的常规对应项相比,具有更高的解码阈值。我们还表明,仅在熵危机之后才发生模场计算的不稳定性,这表明在低温下存在冻结玻璃态相。当没有假定信道噪声的先验知识时(搜索基态),我们发现对常规置信传播的强化策略将使解码阈值提高到比常规置信传播更高的值。该值接近动态转换,在动态转换中,所有本地搜索试探法都无法识别真实消息(代码字或铁磁状态)。在动态过渡之后,具有更高能量密度(比铁磁态大)的亚稳态的数量呈指数增长。当传输信道的噪声水平接近静态转变点时,开始成指数地存在大量共享相同铁磁能的码字。

著录项

  • 作者

    Huang, Haiping;

  • 作者单位
  • 年度 2014
  • 总页数
  • 原文格式 PDF
  • 正文语种 {"code":"en","name":"English","id":9}
  • 中图分类

相似文献

  • 外文文献
  • 中文文献
  • 专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号