tl;dr: You looked at our paper Back-Propagation Learning in Deep
Spike-By-Spike Networks (
https://www.frontiersin.org/articles/10.3389/fncom.2019.00055/full ) and
thought "Interesting idea but I can improve that!" then you may want to
tell us your idea... You could end up working on it as a Ph.D. student /
Postdoc for the next three years in Bremen, Germany.
-------------
The Computational Neuroscience group of Klaus Pawelzik invites
applications for an open Ph.D. student / Postdoc position (E13 TV-L 100%
for 3 years; all genders welcome) in the project "Efficient
Implementation of Spike-by-Spike Neural Networks using Stochastic and
Approximative Techniques". We are looking for a person with a strong
background in mathematics, programming, and machine learning as well as
an intense interest in neuroscience. Someone who is not afraid of
cooperating with engineers, since this is a joint project with a focus
on hardware development.
The overarching goal of our project is to improve the efficiency of
spiking artificial neural networks using hardware and algorithmic
approximation techniques. Specifically, the project focuses on
Spike-by-Spike networks since they offer a balance between computational
requirements and biological-realism which keeps the advantages of the
biological networks while enabling a compact technical realization. To
fully take advantage of the unique features of SbS in terms of
robustness and sparseness, dedicated hardware architectures are required.
You would join in with numerical simulations, theoretical analyses, as
well as through the development of new ideas and approaches for boosting
the performance and capabilities of the Spike-By-Spike model.
Furthermore, you would also work on combining Spike-By-Spike networks
with non spiking deep neuronal networks into hybrid models.
The details can be found at
http://www.neuro.uni-bremen.de/content/open-position-sbs
_______________________________________________
uai mailing list
uai@engr.orst.edu
https://it.engineering.oregonstate.edu/mailman/listinfo/uai
