Tuesday, December 4, 2018

[DMANET] Post Doc position on Graph Kernels

Hello,

We offer a software engineer position in Rouen, France, to work on the
implementation of a C++/Python library. This library aims to become a
standard for the graph community to compute dis/similarities mesures on
graphs.

Contact postdoc-graphkernel@litislab.fr for any questions or check the
offer here : http://pagesperso.litislab.fr/~bgauzere/inge.pdf


Best regards,

Benoit Gaüzère.

Brief Description of the position

Keywords: Graph algorithms, edit distance, discrete optimization, C++,
python.
Duration: 10 months

Context

The edit distance between two graphs G 1 and G 2 can be defined as the
cost induced by the transformation of G 1 into G 2 through elementary
edit operations. This distance, equivalent to the Levenshtein distance
on strings, is very natural and is used by many methods seeking to
predict the properties of objects encoded by graphs.

For several years, the GREYC and LITIS laboratories of Normandy
University have been working on algorithms for computing edit distance
between graphs. These works allowed the proposal of several
state-of--the-art algorithms: see refs[1–3].

As part of the AGAC regional project, the two laboratories are collabo-
rating to pool their expertise in order to design and develop a C++
software library that includes both the methods developed in the two
laboratories and most of the important methods in the field. The
objective here is to create a reference library. In this context, we are
recruiting an engineer for a period of 10 months, whose assignments will
be as follows.

Objectives:

1. Add a number of simple features to the library:
• add some missing methods in collaboration with their authors,
• extend the graph formats supported by the library.
2. Add a python binding
• export the library methods to python,
• Develop import/export functions for the graphs of the python
library (NetworkX/Boost),
• Create a Python pip package to allow an easy and fast use of the
library in Python. Compatibility with sklearn and NetworkX will
be a major asset in the deployment and use of the library.
3. Study the possibility of creating Matlab/Octave binding.

Skills sought

The MSC/engineer level candidate will have good skills in both Python
and C++. Experience in Matlab/Octave as well as graph or pattern
recognition skills would also be appreciated (but not mandatory). Fluent
English is desirable for discussion with researchers from various countries.

Details on the position:

Location: LITIS laboratory in Rouen (Normandy).
Date of desired start January 2019
Duration: 10 months (depending on arrival date)
Salary: about 2200 euros/month net
Contact: postdoc-graphkernel@litislab.fr

Required documents
• Updated CV
• cover letter explaining the candidate's qualifications for the position,
• Letter of support (if applicable)

References
[1] Zeina Abu-Aisheh, Benoit Gaüzère, Sébastien Bougleux, Jean-Yves
Ramel, Luc Brun, Romain Raveaux, Pierre Héroux, and Sébastien Adam.
Graph edit distance contest: Results and future challenges. Pattern
Recognition Letters, 100:96–103, 2017.
[2] Sébastien Bougleux, Luc Brun, Vincenzo Carletti, Pasquale Foggia,
Benoit Gaüzère, and Mario Vento. Graph edit distance as a quadratic
assignment problem. Pattern Recognition Letters, 87:38 – 46, 2017. Ad-
vances in Graph-based Pattern Recognition.
[3] Julien Lerouge, Zeina Abu-Aisheh, Romain Raveaux, Pierre Héroux, and
Sebastien Adam. New binary linear programming formulation to compute
the graph edit distance. Pattern Recognition, 72:254 – 265, December
2017.

**********************************************************
*
* Contributions to be spread via DMANET are submitted to
*
* DMANET@zpr.uni-koeln.de
*
* Replies to a message carried on DMANET should NOT be
* addressed to DMANET but to the original sender. The
* original sender, however, is invited to prepare an
* update of the replies received and to communicate it
* via DMANET.
*
* DISCRETE MATHEMATICS AND ALGORITHMS NETWORK (DMANET)
* http://www.zaik.uni-koeln.de/AFS/publications/dmanet/
*
**********************************************************