ASPICS: Applying Statistical Physics to Inference in Compressed Sensing Image


    This page contains codes, papers and data on the algorithms we have been developing using statistical physics for inference in interdisciplinary applications, with a special focus on compressed sensing.


    Articles

    * Our pionering paper on a new (and powerful) approach to compressed sensing is in open access on Phys. Rev. X 2, 021005 (2012) .
    * A longer version, with many more details and computation, is now published on J. Stat. Mech. (2012) P08009 (and still avaliable on the arxiv).
    * We have also investigated compressed sensing of signal which are only approximatly sparse . This was presented at the 50th Allerton conference.
    * Another interesting case is when the matrix (and the signal) are binary. This has application in the Non-adaptive pooling strategies for detection of rare faulty items .
    * Our more recent works deal with matrix uncertaintly and dictionnary learning and calibration.

    Recent codes

    Our new algorithm for Blind Sensor Calibration can be downloaded here in matlab. The corresponding article has been presented at NIPS 2013.

    The Most recent matlab implementation of our algorithm for compressed sensing can be access on Github: BPCS (Belief Propagation for Compressed Sensing.

    Older Codes and data

    The original MATLAB implementaion can be download here. We would be more than happy to receive comments and suggestions.

    We have also other older implementations: here is a c++ implementation . We have also a a python implementation if you prefer. The fastest implementation is the MATLAB one. Our solver has also been implemented in the KL1p library.

    We have been asked (between other things) to give the data on the spinodal transition that marks the limit of the performance of the EM-BP algorithm for Gauss-Bernoulli signals: here they are. When using a (non-structured) random matrix and a Gauss-Bernoulli signal, this marks the limit of efficiency of sampling algorithm such as Belief-Propagation, and it may be a limit for many other approaches as well. This limit, however, is already beyond the Donoho-Tanner transition. Of course, our seeding strategy allows to break this limit. For reproducible research these are the data and command file we have used in our study, which we give if you want to reproduce our Figure.1 using the c++ code.

    The most recent implementation in MATLAB deals with error-correcting code: download the MATLAB implementation for real-values Error Correcting Code.

    In the news: They spoke about ASPICS!

    * Igor Carron mentioned our work a couple of times in his blog "Nuit Blanche":A stunning development in breaking the Donoho-Tanner phase transition ? and An ASPICS Matlab Implementation.
    * An interview by the French vulgarization Journal, La Recherche Fevrier 2012.


    This is aspics.krzakala.org.
    This page is maintained by Florent Krzakala
    Number of visits:

    joomla site stats