Connectionists: Postdoc position in Bilbao on open-ended network self-organization

[Miguel Aguilera]

Hi all,

We are offering one postdoc position on self-organized network modelling 
at the Basque Center for Applied Mathematics in Bilbao (Basque Country, 
Spain) under the supervision of MIguel Aguilera.

DETAILS: **Postdoctoral Fellow on open-ended, self-organized, 
bio-inspired networks**

Topics: One of the outstanding challenges in modelling living and 
cognitive systems is to capture their ability to continuously adapt and 
develop. Our behavioural responses are not fixed but driven instead by a 
complex ecology, composed of myriads of fluid and inconspicuous 
neurodynamical patterns that have slowly grown on us. AI models like 
neural networks can capture complex behaviours, but often this 
open-ended, liquid and ongoing reconfiguration is incredibly challenging 
to reproduce with static topologies. In contrast, liquid neural networks 
(or 'liquid brains') are a widespread class of networks with a 
particular feature: nodes (which may represent 'neurons' or 'agents' and 
typically all share identical rules) not only process information from 
neighbour nodes, but they also dynamically modify their network 
connections, e.g. by moving in space. These networks show how 
neural-like processing typically associated with static physiological 
networks can also emerge from fluid collective interaction, dynamically 
exploring configuration spaces beyond standard connection weight 
changes. Examples of this kind of model include collective 
decision-making and fluctuations in ant colonies or idiotypic cascades 
in immune networks.

Objective: The aim of this project is to develop a theory of learning in 
liquid brains, focused on two aspects: 1) How do liquid brains learn? 
How is this process different from static neural networks?, and 2) what 
is the adaptive potential of liquid brains when they are embodied as an 
agent in interaction with a changing external environment? Answering 
these questions has the potential to extend the idea of liquid brains 
from a theoretically deep and intriguing concept to a useful tool 
available to the machine learning community. Specifically, liquid brains 
could afford more open-ended, self-improving systems, exploiting fluid 
reconfiguration of nodes as an adaptive dimension which is generally 
unexplored. This could also allow modes of learning that avoid 
catastrophic forgetting, as reconfigurations in the network are based on 
reversible movement patterns. In terms of technology transfer, this 
advances can also have important implications for new paradigms like 
edge computing.

PI in charge: Miguel Aguilera <https://maguilera.net/>

Salary and conditions: The gross annual salary of the Fellowship will be 
29.120€ - 35.360€ according to experience. Additionally, we offer a 
moving allowance up to 2.000€. Should the researcher have a family at 
the time of recruitment: 1. 2.000€ gross in a single payment will be 
offered (you must be married-official register or with children and the 
certificate to prove it must be sent). 2. 1.200€ gross per year/per 
child (up to 2 children) will be offered (the certificate to prove it 
must be sent).

Contract and offer: 1 + 1 years Deadline: 8TH September 2023, 14:00 CET

How to apply: Applications must be submitted on-line at: 
https://joboffers.bcamath.org/apply/ic2023-08-postdoctoral-fellow-in-open-ended-self-organized-bio-inspired-networks

**Scientific Profile Requested**

Requirements: PhD in Physics, Engineering, Computer Science, Maths, 
Artificial Intelligence, Computational Neuroscience and related fields.

Skills and track-record:
* Modelling experience with complex or neural network models, e.g. 
Hopfield networks, Ising models, Boltzmann machines, sandpile models, 
flock/swarm/insect colony models.
* Solid programming skills
* Analytical study of stochastic processes
* Demonstrated ability to work independently and as part of a 
collaborative research team.
* Ability to present and publish research outcomes in spoken (talks) and 
written (papers) form.
* Fluency in spoken and written English.

Scientific Profile:
The preferred candidate will have:
* Strong background in complex systems, neural network modelling, 
biophysics, information theory and/or statistical mechanics
* Interest to work in interdisciplinary research projects

Application and Selection Process Formal Requirements: The selected 
candidate must have applied before the application deadline online at 
the webpage https://joboffers.bcamath.org/ The candidates that do not 
fulfil the mandatory requirements will not be evaluated with respect to 
their scientific profile. Additional documents could be requested during 
the evaluation process so as to check this fulfilment.

Application: Required documents:
* CV
* Letter of interest
* 2 recommendation letters
* Statement of past and proposed future research (2-3 pages)

Evaluation: Based on the provided application documents of each 
candidate, the evaluation committee will evaluate qualitatively: the 
adaption of the previous training and career to the profile offered, the 
recommendation letters, the main results achieved (papers, proceedings, 
etc.), the statement of past and proposed future research and other 
merits; taking in account the alignment of these items to the topic 
offered.

Incorporation: As soon as possible

-- 
Miguel Aguilera | Ikerbasque Research Fellow | BCAM – Basque Center for Applied Mathematics |https://maguilera.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20230901/553b1adf/attachment.html>