Approaching Rate Distortion Bound with Reinforcement Message Passing

• Takumi Fujita

  University of Toyama

• Koji Okino

  University of Toyama

• Tatsuto Murayama

  University of Toyama

抄録

This work reviews the methods for lossy coding of Bernoulli(1/2) source with message passing algorithms/heuristics being imposed some reinforcement assumptions for an optimal convergence. In particular, we consider a kind of decimation technique and presents a tradeoff relation between computational cost and the accuracy of resulting solution for the original problem.

収録刊行物

• IEICE Proceeding Series

  IEICE Proceeding Series 48 C2L-F-2-, 2016-11-27

  The Institute of Electronics, Information and Communication Engineers

キーワード

• Lossy Coding
• Message Passing
• Reinforcement Algorithms
• Cavity Methods

詳細情報

• CRID: 1390564227313789056
• NII論文ID: 230000009348
• DOI: 10.34385/proc.48.c2l-f-2
• ISSN: 21885079
• 本文言語コード: en
• データソース種別:
  • JaLC
  • CiNii Articles
• 抄録ライセンスフラグ: 使用不可