Graph kernels have already been applied to chemoinformatics and are based on structural information encoded within molecular graphs. However, intrinsic properties of atoms and theirs interactions induce some electronic properties which are not explicitly encoded within classic molecular graphs representations. The main purpose of this post doctoral position is to include this information into a new augmented kernel and apply it on some chemoinformatics datasets. The two main steps will be i) to define a new molecular representation encoding local electronic information and ii) to define a new similarity measure as a kernel to compare two molecules encoded in the new proposed representation.
This project will be supervised in close collaboration by LITIS (Rouen, France) and GREYC (Caen, France) laboratories which have a strong expertise on graph kernels for chemoinformatics. The chemical part will be supervised by COBRA laboratory (Rouen, France) which has proposed various atomic descriptors encoding some electronical information. Their expertise will be essential to be able to encode additional information into a new representation for chemical compounds.
Salary: This position will be granted with about 2200 euros/month net salary.
Application domains: machine learning on graphs, chemoinformatics, graph kernels, graph representations
Place: The research will be conducted at LITIS Laboratory (Rouen, France) in Normandy. The LITIS (EA 4108) is affiliated to Normandie University, University of Rouen and INSA Rouen Normandie.
Start date: January/ February 2018
Duration: 20 months according to discussions with the candidate.
Topics: Graph kernels, graph representation, machine learning
You can contact the team via : email@example.com
• PhD or Master in Applied Mathematics or computer science,
• experience in C++, Python or Matlab programming,
• knowledge in kernel methods, graph based approach constitutes an advantage.
Please send the following documents:
• up to date CV,
• Any recommendation letter
• A short document on research experience and interests
Date limite de candidature: 2018-02-01Equipe : DocApp