PhD (MSCA) position: Implementation of neural networks and machine learning-based techniques for multi-core and multi-mode fiber systems

ASTON UNIVERSITY (UK) is pleased to announce ONE position for Early-Stage Researchers (ESRs) within the POST-DIGITAL - the European Training Network on Post-Digital Computing, an Initial Training Network funded by the Marie Skłodowska-Curie Actions Innovative Training Network (H2020-MSCA-ITN-2019-860830). The successful candidate (ESR#7) will work on the ‘Implementation of neural networks and machine learning-based techniques for multi-core and multi-mode fiber systems’. The ESR will be part of of large network of 15 PhD (ESR) students acquireing cutting-edge scientific expertise coupled with a uniquely broad interdisciplinary education, complementary skills, and experience working in both academia and industry.

The position is available from the autumn 2020 for a maximum duration of 36months.

POST-DIGITAL is a pan-European doctoral-level training network that will educate a new generation of highly sought-after engineers and researchers in an emerging field of expertise: advanced optical transmission and signal processing, information theory, nonlinear physics, and machine learning/computer science. The consortium will implement a world-class advanced training programme for 15 Early Stage Researchers, affording them a uniquely broad interdisciplinary education and training which will enable them to design and develop unconventional signal processing elements based on the brain-inspired computing principles. Our 8 academic and 5 Industrial project partners will contribute to a bespoke training program to generate a unique, integrated, learning environment in which graduate students benefit from secondments, industrial links, employability and entrepreneurial skills, leaving them well suited for highly competitive future jobs within academia and industry.

The successful candidate (ESR#7) will work on the ‘Implementation of neural networks and machine learning-based techniques for multi-core and multi-mode fiber systems’. The project relates to new implementations of artificial neural networks and neuromorphic computing principles using multi-core fibres (MCF) and other space division multiplexing technologies for future optical high capacity communication systems. The MCF will act as effective “nodes” of a neural network, where the silica cores will imitate the neurons and synapses, and the signals between nodes (cores) will be transferred through coupling. In particular the project will exploit the effect of nonlinear signal coupling mechanisms to model the nonlinear transfer function of a neuron. Multiple types of interactions between nodes will be also explored to create different architectures of multi-layer neural networks. The successful candidate will aim to optimize the optical field and fiber parameters and various designs of MCF to achieve the best performance. The new system designs will be studied, first, theoretically, and then validated and investigated experimentally.

The successful candidate will join an established group working on advanced space and mode division multiplexing communication systems incorporating the machine-learning based signal processing elements under the general leadership of Prof. Sergei Turitsyn and with the co-supervision of a senior AITP member.

The ideal candidate

You will have have a Master of Science degree (or expected to be awarded one soon) in Physics, Electrical Engineering, Applied Mathematics, or equivalent. You will have exceptional skills in the areas related to electronic engineering, physical science, nonlinear optics, photonics, applied mathematics and machine learning. Preferred skill requirements include experience of scientific programming, optical waveguide theory and communication theory. Knowledge in theory of linear algebra, mathematical physics, and numerical analysis are an asset. You will have a strong drive to carry out the cutting edge research for a doctoral degree.