[Comp-neuro] Post-Doc Position on EU Project

taymore2002 at aol.co.uk taymore2002 at aol.co.uk
Thu Nov 11 11:47:05 CET 2010


Post-Doctoral Job Application for the EU Project DARWIN
A Post-Doctoral position will be open starting on Feb 1, 2011 for 3 
years in the EU Project DARWIN (DARWIN denotes“Dextrous Assembler Robot 
Working with Embodied Intelligence”, with the 6 partners:
1) Profactor GMBH, Austria,
2) Foundation for Research and Technology, Heraklion, Greece,
3) King’s College London, United Kingdom,
4) Fondazione Istituto Italia di Tecnologia, Italy,
5) Ceske Vysoke Uceni Technice v Praze, Czech Republic,
6) Ε. Κasderidis & Sia Ε.Ε., Thessaloniki, Greece)
Targeting both assembly and service industry the DARWIN project aims to 
develop an ‘acting, learning and reasoning’ assembler robot that will 
ultimately be capable of assembling complex objects from its 
constituent parts. The work of DARWIN is based mainly on the earlier EU 
Projects GNOSYS & MATHESIS, in which neural architectures capable of 
internal reasoning and observational learning were developed, guided by 
the mechanisms observed in the human brain through brain imaging and 
psychological experiments (JG Taylor et al, Image and Vision Computing, 
2009, 27(11):1641-1657). The work to be carried out in DARWIN by KCL 
will be guided by Professor JG Taylor and his colleagues (Professor L 
Seneviratne, Drs Althoefer, Nanayaar & Sprattling) in the Information 
Technology Department of KCL. There will be considerable co-ordination 
of this work with the partners across Europe and meetings at the 
various partners will take place several times every year.
The detailed tasks of KCL in DARWIN are:
Neural coding of objects and attention control: To develop a ‘high 
resolution (visual) perception’ system, where object classes will have 
representations based on neural activations. A hierarchical visual 
system will be developed, guided by corresponding brain structures, 
which will build different representations ranging from low level 
features to more complex ones such as complex silhouettes, colour and 
texture to top level ones, such as object classes. The system will 
integrate the ‘what’ and ‘where’ of visual processing. It will 
implement internally a mechanism of attention in the sense of 
strengthening the activation of attended objects by learnt feedback.
Action Affordances: To learn affordances of objects, so that these can 
be used in the process of planning and making actions by the 
development of internal models (in collaboration with Pietro Morasso in 
Italy). At the same time the interaction of the sensory (visual and 
somato-sensory) systems is to be developed in order to make actions 
effective.
Skills required: PhD in neuro-computation or robotics, with an interest 
in the workings of the brain and mind, and if possible modelling of 
visual or action processing. Programming skills are assumed at the 
usual level of C/C++/etc.
Salary etc. At standard UK Post-Doc rates, depending on age and 
experience.
Reference
Taylor JG, Hartley M, Taylor NR, Panchev C & Stathis Kasderidis S, 
(2009) Hierarchical Attention Based Visual System: Perception, Action 
and Reasoning by Brain Guidance. Image and Vision Computing 
27(11):1641-1657



More information about the Comp-neuro mailing list