Statistical Physics and Interdisciplinary Applications

Attachment	Size
gf-rbp.tgz	24.36 KB

This code implements a novel data compression technique for binary symmetric sources based on the cavity method over GF(q), the Galois Field of order q. We present a scheme of low complexity and near-optimal empirical performance. The compression step is based on a reduction of a sparse low-density parity-check code over GF(q) and is done through the so-called reinforced belief-propagation equations. These reduced codes appear to have a nontrivial geometrical modification of the space of codewords, which makes such compression computationally feasible.

Braunstein A, Kayhan F, Zecchina R. Efficient LDPC Codes over GF(q) for Lossy Data Compression. In: IEEE International Symposium on Information Theory, 2009. ISIT 2009. Seul, Korea; 2009.

Attachment	Size
ISIT2009.pdf	814.41 KB

Ciliberti S, Mézard M, Zecchina R. Message-Passing Algorithms for Non-Linear Nodes and Data Compression. ComPlexUs. 2006;3:58-65.

Zecchina R. Statistical physics, optimization and source coding. Pramana. 2005;64:1161-73.

Battaglia D, Braunstein A, Chavas J, Zecchina R. Source coding by efficient selection of ground-state clusters. Phys Rev E. Stat Nonlin Soft Matter Phys. 2005;72:015103.