In the COATI joint team between INRIA Sophia Antipolis and the I3S Lab at Université Côte d'Azur, we are looking for a postdoc to join a project which aims at investigating algorithmic procedures for improving the speed with which neural networks can be evaluated, and at shedding light on how much information about the computational problem at hand can be encoded within neural networks of small size. In particular, the project focuses on the design of sparsification algorithms that, given a pre-existing network, modify its structure by exploiting the high level of redundancy that is typical of modern architectures, and produce a network with a reduced number of neurons and a topology which is adapted for novel techniques which speed network evaluation up. The ideal candidate has a solid background in algorithms and/or mathematics, values developing scientific software and is interested in applying algorithmic techniques to the problem of sparsifying artificial neural networks and related problems.
Prior knowledge on artificial neural networks is not required. Implementations and experiments in the project are being developed using the Julia programming language, therefore a desire for getting rapidly familiar with it is expected.
The postdoc is expected to conduct original scientific research and develop software within the scope of the project, in collaboration with Emanuele Natale and Laurent Viennot.
The candidate will also have the opportunity to periodically visit the IRIF Laboratory in Paris. For more information and to apply for the position, please visit the website https://recrutement.inria.fr/public/classic/fr/offres/2022-04607
Best regards,
Emanuele Natale
**********************************************************
*
* 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/
*
**********************************************************
Posts
