- Entry level
- No Education
- Salary to negotiate
The position holder will be a member of the group of Herbert Jaeger (https://www.rug.nl/staff/h.jaeger), who recently joined the University of Groningen. The position is funded through a personal startup fund, not by a third-party project, and offers great freedom for the position holder to define his/her own research agenda as long as it contributes to a mathematical / conceptual understanding of neuromorphic computing.
The local scientific context is the Research Center for Cognitive Systems and Materials (CogniGron, https://www.rug.nl/research/fse/cognitive-systems-and-materials/?lang=en). This is a recent, very substantial research initiative (10 new professorships, 30 PhD positions) of the University of Groningen. Jointly managed by the Bernoulli Institute (mathematics, computer science, AI) and the Zernike Institute for Advanced Materials, the mission of CogniGron is to "to develop materials-centered systems paradigms for cognitive computing based on modelling and learning at all levels: from materials that can learn to devices, circuits and algorithms". This is a decidedly interdisciplinary enterprise at the interfaces between the materials sciences, microchip technology, computational neuroscience, cognitive science, AI, computer science and mathematics. Together with previously existing research groups at the university, the new ten CogniGron research groups will form a scientific microcosm of 20+ research teams.
Within this context, the group of Herbert Jaeger will work toward rigorous theoretical foundations of "neuromorphic computing". At present, this is only an umbrella term which covers a diversity of largely unconnected and limited models and methods. A guiding idea in this emerging field is to argue that (A) novel, energy-efficient, non-digital microchips share many properties with biological brains (such as extremely high dimensionality, spatial and modular organisation, strong nonlinearity, stochasticity, low numerical resolution, parameter drift, aging, always-on real-time operation); (B) biological brains function exceedingly well; (C) hence, it should be possible to realize high-performing computing in neuromorphic microchips by "learning from the brain". However, the neurosciences do not yet deliver a comprehensive theory of how the brain "computes". The research mission of Jaeger's group is to develop new formal, conceptual and algorithmic tools which can fill some of the gaps in our theoretical understanding. This may lead to the invention of novel mathematical description languages which unify elements of symbolic information processing (logic formalisms, Turing computability) and analog information processing (nonlinear dynamics, signal processing and control theory).
- ms project