Doc Bernds FutureLab

Can we Copy the Brain?

The brain is one of the most fascinating parts of the body and despite all the existing knowledge, many questions still remain. That is why I have chosen the motto "Can we copy the brain" for the Xyna Conference 2017. A central question that has been with us for a long time and that we are also currently dealing with. Particularly relevant are the two approaches of Deep Learning Neural Networks and Neuromorphic Computing. In particular, the latter approach, in execution with Spiking Neural Networks, offers an interesting approach to non-digital algorithmic intelligence. Representing Deep Learning Networks, I was able to attract one of the most famous researchers in artificial intelligence, Prof. Dr. Schmidhuber, to speak at our 2017 Xyna conference. His talk gave the guests an exciting insight about the future development of artificial intelligence and into his research work. In doing so, he came to talk about his vision in which outer space would be conquered by machines with artificial intelligence.  His machine-centered and posthuman vision of the future was countered by philosopher Prof. Dr. Gehring, who presented a philosophical and ethical view of the topic. If you want to read even more about our this conference, click here.

Since 2017, a lot has happened in the field of artificial intelligence. Therefore, I am currently rephrasing the question "Can we copy the brain" to "Can we support the brain". Unlike the first, the second question can already be answered in the affirmative.

The approach of a data-driven algorithmic intelligence, as offered by AI approaches such as DLN, implies that the right data can be provided in sufficient quantity for their training. While this may still be comparatively easy for cat image recognition, it is not trivial for more complex problems. The key is to define the right task and to acquire the right data with the necessary quantity and quality. This shifts the task from knowledge-based model building to data acquisition, or as I would describe it as a physicist: the determination of the relevant parameters with the help of the corresponding measurement methods.

I am currently working on the question of how routing and beamforming could be implemented in 5G and XG mobile communications using AI. In this context, I refer to the article by F. Restuccia and T. Melodia in IEEE Communications Vol. 58 No. 10 p.58.