Package: sponsorship-requests Severity: wishlist Dear mentors,
We are looking for a sponsor for the "toulbar2" software, in the debian-science project. A current version of its git repository is available on alioth, in the debian- science/pkg-toulbar2 git repository. Toulbar2 is an open source exact discrete optimization software targeted at solving optimization problems that are described as "Graphical Models" including Cost Function Networks, aka Weighted cConstraint Satisfaction Problems, Markov Random Fields (MAP/MRF), Bayesian Nets (MPE), Weighted MaxSAT, pre linkage files and Quadratic Pseudo Boolean Optimization problems. On such problems, toulbar2 is often more efficient than expensive commercial ILP (Integer Linear Programming) solvers. Toulbar2 has won international solver competitions: the Weighted CSP competition (first, in 2007 and 2008), the Uncertainly in AI challenge (first, in 2010 and 2014) and the Pascal Inference challenge (second, in 2011). It has been used in several scientific publications in machine learning, theoretical computer science, statistical physics, genetics and structural biology. It has been developed for more than 10 years on our FusionForge server, hosted by the MIA (Applied mathematics and Computer Science) Departement of INRA, offering a stable and reliable environment. The forge also hosts the associated costfunctionlib benchmark library: https://mulcyber.toulouse.inra.fr/projects/toulbar2/ https://mulcyber.toulouse.inra.fr/projects/costfunctionlib As a preliminary test, a prerelease of a debian source package can be obtained at https://launchpad.net/~thomas-schiex/+archive/ubuntu/toulbar2 -- System Information: Debian Release: stretch/sid APT prefers xenial-updates APT policy: (500, 'xenial-updates'), (500, 'xenial-security'), (500, 'xenial'), (100, 'xenial-backports') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.4.0-14-generic (SMP w/4 CPU cores) Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system)
