TY - GEN

T1 - Multi-bit flipping algorithms with probabilistic gradient descent

AU - Vasic, Bane

AU - Ivanis, Predrag

AU - Brkic, Srdan

N1 - Funding Information:
This work was supported by the National Science Foundation under Grants ECCS-1500170 and CCF-1314147, IndoUS Science and Technology Forum (ruSSTF) through the Joint Networked Center for Data Storage Research (JC-16-2014-US)and by the Serbian Ministry of Science under Project TR32028.

PY - 2017/8/30

Y1 - 2017/8/30

N2 - A new class of bit flipping algorithms for low-density parity-check codes over the binary symmetric channel is proposed. The algorithms employ multiple bits at a variable node to represent its reliability, and multiple bits a check node to capture the sequence of its syndrome values. The check node update function thus requires a simple bit-shift operation, while the variable node updates require a nonlinear Boolean function. This class of multi-bit flipping (MBF) algorithms is enhanced by the probabilistic gradient descent (PGD) algorithm. The gradient descent algorithm minimizes the variable node energy function which, in addition to the classical term which quantifies the discrepancy between the variable estimate and channel value, also involves an additive term defined as a weighted sum of neighboring check node states. Only the variable nodes with the maximal value of energy are eligible for updating, but the updates are not done by default but probabilistically. The resulting probabilistic gradient descent multi-bit flipping PGD-MBF algorithm combined with rewinding improves the codeword probability of error while keeping the complexity lower than that of the state-of-the-art algorithms of comparable throughput.

AB - A new class of bit flipping algorithms for low-density parity-check codes over the binary symmetric channel is proposed. The algorithms employ multiple bits at a variable node to represent its reliability, and multiple bits a check node to capture the sequence of its syndrome values. The check node update function thus requires a simple bit-shift operation, while the variable node updates require a nonlinear Boolean function. This class of multi-bit flipping (MBF) algorithms is enhanced by the probabilistic gradient descent (PGD) algorithm. The gradient descent algorithm minimizes the variable node energy function which, in addition to the classical term which quantifies the discrepancy between the variable estimate and channel value, also involves an additive term defined as a weighted sum of neighboring check node states. Only the variable nodes with the maximal value of energy are eligible for updating, but the updates are not done by default but probabilistically. The resulting probabilistic gradient descent multi-bit flipping PGD-MBF algorithm combined with rewinding improves the codeword probability of error while keeping the complexity lower than that of the state-of-the-art algorithms of comparable throughput.

UR - http://www.scopus.com/inward/record.url?scp=85031033982&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85031033982&partnerID=8YFLogxK

U2 - 10.1109/ITA.2017.8023480

DO - 10.1109/ITA.2017.8023480

M3 - Conference contribution

AN - SCOPUS:85031033982

T3 - 2017 Information Theory and Applications Workshop, ITA 2017

BT - 2017 Information Theory and Applications Workshop, ITA 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2017 Information Theory and Applications Workshop, ITA 2017

Y2 - 12 February 2017 through 17 February 2017

ER -