Among the research activities carried out for the Fun-COMP project, the simulation methodology is an invaluable tool to move steps towards a better understanding of the devices mechanisms and behaviour.

Simulation, or modeling, usually consists of the mathematical replica of a device, over which the known laws of physics are applied, and whose results are tested against the experimental data. The general purpose is dual: expand the knowledge over the physical mechanisms and material properties governing the device behaviour, and use this knowledge to infer any possible improvement without the immediate need of an experimental exploration.

Indeed, simulation is quite a broad term, encasing several different approaches and aims. For the Fun-COMP research we developed several different methodologies, two of these dedicated to tackle only the unit cell behavior. One is the Finite Element model, defined to explore and test the unit cell physical laws, and the second one is the Behavioral model, employed to predict larger scales integration.

Further information on the simulation methods is contained in this report.

“Researchers take a step towards light-based brain-like computing”

Congratulations to Johannes, Nathan, David, Harish and Wolfram, for their publication in Nature of the development of a brain-like optical computing chip.

A press release is available at this link, and the article can be found at this page.

There is also a nice “Nature News and Views” article highlighting the achievements of the paper at this page.

Many thanks to Dr. Abu Sebastian and Dr. Simon Stringer!

During the 1st of our programmed cross-disciplinary training session, they delivered two extremely interesting talks.

Dr. Sebastian is Principal Research Staff Member at IBM Zurich, particularly focused on neuromorphic and in-memory computing research. He explored the fundamentals and applications of phase-change memory devices.

Dr. Stringer leads the Oxford Centre for Theoretical Neuroscience and Artificial Intelligence, based within the Department of Experimental Psychology. He provided a picture of the modelling techniques in brain science, in particular tackling the “binding problem” embedded in spiking neural network.

Andrew Katumba and Peter Bienstman form Imec published their most recent work on Journal of Lightwave Technology, “A Neuromorphic Silicon Photonics Nonlinear Equalizer for Optical Communications with Intensity Modulation and Direct Detection” (link)