Researchers in Switzerland have developed an optical computer for machine learning that consumes 100 times less power than today’s GPU-based systems.
The team at EPFL in Lausanne has developed an optical computer for machine learning using optical fibres. The Scalable Optical Learning Operator (SOLO) can recognize and classify information formatted as two dimensional images. This was compared to current neural networks on 3,000 X-rays of Covid-19 patients and showed a dramatic lower power consumption.
“Light transmits information without any physical interference from cables. That’s the core advantage of optical technology when it comes to transferring data," said Demetri Psaltis, head of EPFL’s Optics Laboratory within the School of Engineering. “To take artificial intelligence as an example, many AI programs require accelerators to carry out rapid calculations using minimal power. For now, while optical technology could theoretically meet that need, it has not yet reached the applied stage – despite a half-century of research. That’s because optical computing and decision-making do not yet save either time or energy.”
Designing optical computing devices remains a challenge. Although the computations are performed rapidly , the obstacle comes in transferring the results to memory at that same speed and in an energy-efficient manner.
Engineers at Psaltis’ lab, along with colleges at Christophe Moser’s Laboratory of Applied Photonic Devices, also within the School of Engineering, have developed the SOLO machine learning method, published in Nature Computational Science.
- Optical neural network accelerator for machine learning
- 11 TOPS photonic convolutional neural network chip
- Optical processor speeds compute for next-gen AI
- imec teams for quantum photonic circuits
- Experimental optical chip works like a brain
The machine learning uses the combination of the linear and nonlinear parts of the optical system in a shared volume confined in a multimode fibre (MMF). The principal advantage of this approach is the combination of the three-dimensional connectivity of optics with the long interaction length and lateral confinement in the fibre, using