Polaritonic Neuromorphic Computing Outperforms Linear Classifiers
| Autor: | Dario Ballarini, Lorenzo Dominici, Milena De Giorgi, Giovanni Lerario, Giuseppe Gigli, Michał Matuszewski, Antonio Gianfrate, Timothy Chi Hin Liew, Vincenzo Ardizzone, Daniele Sanvitto, Riccardo Panico, Sanjib Ghosh, Andrzej Opala |
|---|---|
| Přispěvatelé: | School of Physical and Mathematical Sciences, Ballarini, D., Gianfrate, A., Panico, R., Opala, A., Ghosh, S., Dominici, L., Ardizzone, V., De Giorgi, M., Lerario, G., Gigli, G., Liew, T. C. H., Matuszewski, M., Sanvitto, D. |
| Rok vydání: | 2020 |
| Předmět: |
FOS: Computer and information sciences
Computer science Big data Computer Science - Emerging Technologies FOS: Physical sciences Bioengineering Linear classifier 02 engineering and technology Bottleneck symbols.namesake Optical Microcavities Neuromorphic computing General Materials Science Physics::Optics and light [Science] Exciton-polaritons Reservoir computing Artificial neural network Exciton-polariton Optical microcavitie business.industry Mechanical Engineering General Chemistry Disordered Systems and Neural Networks (cond-mat.dis-nn) Condensed Matter - Disordered Systems and Neural Networks 021001 nanoscience & nanotechnology Condensed Matter Physics Emerging Technologies (cs.ET) Neuromorphic engineering Computer engineering Semiconductors Quantum Gases (cond-mat.quant-gas) symbols 0210 nano-technology business Condensed Matter - Quantum Gases MNIST database Von Neumann architecture |
| Zdroj: | Nano letters. 20(5) |
| ISSN: | 1530-6992 |
| Popis: | Machine learning software applications are ubiquitous in many fields of science and society for their outstanding capability to solve computationally vast problems like the recognition of patterns and regularities in big data sets. In spite of these impressive achievements, such processors are still based on the so-called von Neumann architecture, which is a bottleneck for faster and power-efficient neuromorphic computation. Therefore, one of the main goals of research is to conceive physical realizations of artificial neural networks capable of performing fully parallel and ultrafast operations. Here we show that lattices of exciton-polariton condensates accomplish neuromorphic computing with outstanding accuracy thanks to their high optical nonlinearity. We demonstrate that our neural network significantly increases the recognition efficiency compared with the linear classification algorithms on one of the most widely used benchmarks, the MNIST problem, showing a concrete advantage from the integration of optical systems in neural network architectures. Accepted version ERC “ElecOpteR” Grant 780757 Singapore, MOE2017-T2-1-001 Singapore, MOE2018-T2-02-068 Poland, 2016/22/E/ST3/ 00045 Poland, 2017/25/Z/ST3/03032 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|