Electronic-photonic Architectures for Brain-inspired Computing

feed news

HYBRAIN and the role of photonic technologies in the EU innovation landscape

Moore’s law states that the number of transistors in a microprocessor doubles roughly every two years, increasing chip power while reducing energy usage. Moore’s foresightful vision has guided the astounding advancements made in the field of computing for half a century. When transistors get smaller in order to fit more of them into a given volume, their speed increases and their energy consumption decreases.

Despite that, the demise of Moore’s law is an inevitable consequence of physics: this reduction cannot continue indefinitely. Although adding more transistors can always result in more powerful hardware, doing so will slow down processing and cause the chips to become hotter.

Therefore, in order to advance computing technology without adding transistors, the hardware must be changed. The next stage in the development of computing is photonics.

Photonics vs electronic computing

Photonics has several advantages over today’s electronic hardware, including increased bandwidth, processing power, and storage space. A photonic hardware manipulates, stores, and transmits data using light rather than electricity (photons rather than electrons), which is one of the factors that led to the evolution of photonics because it knits the rate at which data can be exchanged.

Photonics hardware can potentially overcome the most pressing shortcomings of electronic hardware:

  • Light can travel through free space without the need for wires or fibres, and photons can travel through each other without alteration.
  • Energy losses from light traversing free space are negligible, allowing highly energy-efficient devices.
  • While photonics hardware can be slowed down by electro-optic switches, some photonics computations can actually be carried out at the speed of light.

Given its advantages and disadvantages, the general uses of photonics computing are not in displacing general-purpose electronic hardware but rather in more specialised fields where its advantages are greatest. The interconnection with traditional electronic computing chips or boards is the most visible example of this.

In fact, for machines with multiple processors, photonics has the potential to greatly improve connection and cut down on communication times.

Hybrid photonics applications

Photonics computing for specialised workflows already has its own niches, especially in data centres and more recently in machine learning and Artificial Intelligence (AI) applications.

In particular, neural network development is a key application area. These are self-learning networks for pattern recognition and image processing that were initially developed to mimic human and animal neural processing. Since photonics approaches are highly parallel and depend on each unit interacting with every other one, they are perfect for neural networks.

Photonics data processing can be done in parallel (which means that they are able to carry out several tasks at once) much more easily and cheaply than electronic data processing, where the instructions are instead carried out one at a time.

The EU photonic innovation agenda and the role of HYBRAIN

Given its innovation potential, photonic technologies are of paramount importance for Europe. In fact, the European Commission is working to ensure that individuals and companies can reap the full benefits of this technology.

Through a public-private partnership (PPP) and European Union funded scientific research projects, such as HYBRAIN, Europe will try to cement its industrial leadership in photonic technologies that will result in a shorter time to market for innovative products, allowing for an overall benefit in terms of jobs creation and growth in Europe, more generally aiming to:

  • Stimulate investment into private sector research, development and innovation.
  • Promote research and innovation with a business-driven agenda.
  • Accelerate the development of photonic technologies.

The rapid development of artificial intelligence (AI) facilitates various applications from all areas but also poses great challenges in its hardware implementation in terms of speed and energy because of the explosive growth of data shared and produced.

HYBRAIN, with its edge-computing technology, provides a distinctive perspective to address this bottleneck by harnessing the unique properties of photonic technologies, including broad bandwidth, low latency, and high energy efficiency.

Therefore, the project will contribute to fostering the creation of a shift in how we build computing hardware: the future is hybrid photonic-electronic technologies.

Scroll down to subscribe to our newsletter to stay up to date with our research activities and results.