The modern information and communication technology (ICT) world is ruled by an exploding generation of data due to growth of Internet applications and skyrocketing proliferation of a number of interconnected on-line broadband services (clouds, Facebook, Google, YouTube, on-demand HD video streams, on-line business analytics and content sharing, telepresence, and so on), smartphones, industrial sensor networks, and other applications. The analysis of this data demands a massive increase in processing power, memory and communication bandwidth.

In the past, this ever-growing demand could be provided by the exponential increase in the performance of digital data storage and processing technologies – «Moore’s law». However, technological progress rates of digital technologies are flattening out and in a decade or two will hit ultimate physical ceilings («end of Moore’s law»). This growing gap between what is demanded and what can be delivered becomes exacerbated by the increasingly inacceptable energy consumption of our global ICT infrastructure. In this situation, both academic and industrial ICT research has started to investigate and invest in so-called «unconventional», nature-inspired approaches to computing which radically depart from the classical digital paradigm.

Two major goals in this pursuit are to exploit the potential speed and energy advantages that computing with light offers over computing with electrons, and the advantage that analog computing offers over digital computing based on parallel processing and the adaptive intelligence of biological brains, captured in novel artificial neural network designs. In POST-DIGITAL we put a special focus on analog processing of optical signals in neural network architectures because of the speed and bandwidth provided by photonics. The future economic growth in Europe requires engineers and researchers capable to design, develop and implement new information technologies to support explosive data-driven transformation of economy, public and government activities.

Two major application domains at least would crucially benefit from harnessing the powers of neurons and photons. First, the world’s backbone communication infrastructure could be salvaged from data suffocation by the orders-of-magnitude higher bandwidths and similarly orders-of-magnitude lower energy costs affordable, in principle, by photonics. Second, end-user mobile systems, prosthetic implants, «Internet of Things» network of devices and a host of other end-of-line products could be designed to be inherently adaptive, robust, and life-long learning and minimum energy consumption. However, the theoretical and engineering challenges posed by new computational principles and unconventional physical substrates are by no means resolved, and a new generation of researchers and engineers needs to be sent out on a new trajectory that is likewise daring, demanding, and promising.

 

POST-DIGITAL brings together fourteen leading academic and industrial players (including IBM, Thales and three highly reputed SMEs) in optical and neuromorphic computing, already connected in long-standing collaboration networks. These groups will now join forces in order to train through ground breaking research, complimentary skills training and industrial secondments fifteen high calibre early-stage researchers to become backbone innovators in a technology domain that is crucial for Europe’s sustained competitiveness in ICT scenarios facing a dramatic pressure for fundamental re-thinking.

 

Leave a Reply