Changes between Version 12 and Version 13 of WikiStart


Ignore:
Timestamp:
Feb 12, 2009 2:33:01 PM (11 years ago)
Author:
katyas
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • WikiStart

    v12 v13  
    1 = Welcome to the SVM-QP home page =
     1= Welcome to the OPTIML home page =
    22
    33== Introduction ==
     4
     5OPTIML stands for Optimization methods in Machine Learning. Right now this page contains three subprojects:
    46
    57SVM-QP ('''S'''upport '''V'''ector '''M'''achines '''Q'''uadratic '''P'''rogramming solver) is a software package that solves 2-norm soft margin support vector machine [http://www.kernel-machines.org/ classification problem].  The problem is formulated as a convex QP of the following form
     
    1315where {{{ e }}} is the vector of ones, {{{a}}} is a vector of lables (1 or -1) of the data, and {{{C}}} is the panalty parameter associated with  the violation of the margin constraint.
    1416
    15 SVM-QP is designed for  large-scale SVM problems. The underlying algorithm is an active set method for convext QPs. [http://www.research.ibm.com/people/k/katya/incas.pdf/ Here ] is the paper describing the algorithm and containing computational comparison with [http://svmlight.joachims.org/ SVMlight]. The software also includes an interior point
     171. SVM-QP is designed for  large-scale SVM problems. The underlying algorithm is an active set method for convext QPs. [http://www.research.ibm.com/people/k/katya/incas.pdf/ Here ] is the paper describing the algorithm and containing computational comparison with [http://svmlight.joachims.org/ SVMlight]. The software also includes an interior point
    1618SVM solver which is designed for problems where Kernel matrix is (approximately) low rank. The program constructs the low
    1719rank approximation and solves the approximate problem by the interior point method. The approximate solution can then be passed
    1820to the active set solver as the warm start. The underlying algorithms are described [http://www.ai.mit.edu/projects/jmlr/papers/volume2/fine01a/fine01a.pdf here].
    1921
    20 The currently available versions is a Fortran 77 code which is designed to trade memory for
     22This versions is in Fortran 77 code and is designed to trade memory for
    2123efficiency. Currently, this is the most time efficient version of SVM-QP.
    22 A memory saving versions, which is somewhat slower will be available soon.
    23 A beta versions can be obtained by contacting [http://www.research.ibm.com/people/k/katya/ Katya Scheinberg].
     24A memory saving versions, which is somewhat slower is available in the C++ version (see SVMPath).
     25In case of questions,  contact [http://www.research.ibm.com/people/k/katya/ Katya Scheinberg].
    2426The current version is desinged to ba called as a subroutine where the burden
    2527of parsing the data and arranging it into appropriate data structures lies with the user.
    2628
    27 The C++ version of the software is under development by [http://web.mit.edu/belloni/www/ Alexandre Belloni]
    28 and will be available in the near future. Additionally the C++ version will include the ability compute a path of optimal solutions for any given range of parameter {{{C}}}.
    2929
     302. SVMPath is the C++ extension of SVM-QP.  Additionally to solving the SVM problem for a given value of parameter C, the C++ version includes the ability compute a path of optimal
     31solutions for any given range of parameter {{{C}}}. This version is strictly in the development stage right now. The manual and other supporting materials will be coming soon.
     32Please check back or contact the authors.
     33
     343. SINCO ('''S'''parse '''IN'''verse '''CO'''variance) is the C++ software with a Matlab interface which solve the sparse inverse covariance selection problem
     35{{{
     36   max K ln det{C} -tr(AC) -\lambda||S.*C||_1
     37C in S^{p x p}
     38}}}
     39where {{{K, \lambda}}} are positive scalars, {{{A}}} is a {{{p x p}}} symmetric matrix, {{{S}}} is a {{{p x p}}} nonnegative matrix and {{{C}}} is the unknown {{{p x p}}}
     40symmetric positive definite matrix. This software is written in C++ and has a Matlab interface which is provided. A brief description of how to use the code is available via
     41comments. A more detailed manual and the algorithm description is forthcoming.
    3042-------------
    3143
    3244== Background ==
    3345
    34 SVM-QP is released as open source code under the [http://www.opensource.org/licenses/cpl.php Common Public License (CPL)].  It is available from the [http://www.coin-or.org/ COIN-OR initiative].  [http://www.research.ibm.com/people/k/katya/ Katya Scheinberg] is the COIN project leader for SVM-QP.
     46OPTIML is released as open source code under the [http://www.opensource.org/licenses/cpl.php Common Public License (CPL)].  It is available from the [http://www.coin-or.org/ COIN-OR initiative].  [http://www.research.ibm.com/people/k/katya/ Katya Scheinberg] is the COIN project leader for OPTIML.
    3547
    36 You can obtain the SVM-QP code either via subversion or in form of nightly generated tarballs.  To get the tarball, go to the COIN [http://www.coin-or.org/download/source/SVM-QP tarball directory], and look for a file like svmqp-3.x.x.tar.gz.
     48You can obtain any of the softwares available in OPTIML code either via subversion or in form of nightly generated tarballs.  To get the tarball, go to the COIN [http://www.coin-or.org/download/source/SVM-QP tarball directory], and look for a file like svmqp-3.x.x.tar.gz.
    3749
    38 Individual files can also be obtained from the svn web interface (see the "Browse Source" button above).  The SVM-QP distribution can be used to generate a library that can be linked to one's own C++, C, or Fortran code. SVM-QP
    39 can be used on Linux/UNIX platforms and Windows.
     50Individual files can also be obtained from the svn web interface (see the "Browse Source" button above).  The SVM-QP and SVMPath distributions can be used to generate a library that can be linked to one's own C++, C, or Fortran code. SVM-QP and SVMPath
     51can be used on Linux/UNIX platforms and Windows. SINCO can be used from C++ and from Matlab via the available mex interface.
    4052
    41 As open source software, the source code for SVM-QP is provided without charge.  You are free to use it, also for commercial purposes.  You are also free to modify the source code (with the restriction that you need to make your changes public if you decide to distribute your version in any way, e.g. as an executable); for details see the CPL license.  And we are certainly very keen on feedback from users, including contributions!
     53As open source software, the source code for OPTIML projects is provided without charge.  You are free to use it, also for commercial purposes.  You are also free to modify the source code (with the restriction that you need to make your changes public if you decide to distribute your version in any way, e.g. as an executable); for details see the CPL license.  And we are certainly very keen on feedback from users, including contributions!
    4254
    4355In order to compile SVM-QP, certain third party code is required (such as some linear algebra routines).  Those are available under different conditions/licenses.
    4456
    45 We provide this program in the hope that it may be useful to others, and we would very much like to hear about your experience with it.  If you found it helpful and are using it within our software, we encourage you to add your feedback to these wiki-based webpages, [wiki:SuccessStories Success Stories].
     57We provide these programs in the hope that it may be useful to others, and we would very much like to hear about your experience with it.  If you found it helpful and are using it within our software, we encourage you to add your feedback to these wiki-based webpages, [wiki:SuccessStories Success Stories].
    4658-------------
    4759