An artificial polariton neuron as a step towards a photonic system that mimics how the human brain works

Scientists from the Faculty of Physics at the University of Warsaw and the Polish Academy of Sciences used photons to create a spiking neuron, ie the basic element of the future photonic neural network processor. The so-called neuromorphic devices, systems that mimic the behavior of the biological brain, that researchers are working on are the future of artificial intelligence, as they enable much faster and more effective information processing. We can read about the results of their work in the current “Laser and Photonics Review”.

The mammalian brain is one of the most complex and efficient systems on earth. As early as the 1990s, neurobiologists showed that a single area of ​​the macaque cortex could analyze and classify visual patterns in just 30 milliseconds, despite the fact that each of the neurons involved in this process sends fewer than three messages in the form of electrical impulses. This is made possible by a large number of synapses – the connections between neurons – in the neural network of the macaque brain.

The human brain is part of an even more powerful machine. It consists of 100 billion neurons, each of which makes several thousand connections to other nerve cells on average. This creates a neural network of approximately 100 trillion connections, thanks to which our brain is able to recognize, infer and control movements simultaneously – it performs trillions of operations per second while using only 20-25 watts of power. In comparison, traditional processors use ten times the power just to recognize a thousand different types of objects. This amazing difference and the extraordinary performance of the brain are due to the biochemistry of neurons, the architecture of neural connections and the biophysics of neural computational algorithms, among other things.

Society’s thirst for information is constantly growing, so we have to process this information faster and more comprehensively. Traditional computing systems may not be able to meet the growing demand for more computing power while increasing energy efficiency. The solution to the problem could be the so-called neuromorphic devices, which mimic the actions of the biological brain. They are the future of artificial intelligence as they enable much faster and more effective processing of information in tasks such as image recognition.

Scientists from the Faculty of Physics at the University of Warsaw and the Polish Academy of Sciences, in an article published in Laser and Photonics Review, proposed using photons in a way that allows the creation of spiking neural networks. Krzysztof Tyszka from the Faculty of Physics at the University of Warsaw, the first author of the work, emphasizes that photonic systems ensure light-speed communication, low losses and low energy consumption. The advantage of photons is that their propagation occurs with practically no loss of energy. – Since they interact relatively weakly, it is unfortunately difficult to perform calculations with them analogous to electronic systems – adds the scientist.

– In our research we propose a solution where photons interact strongly with particles of very low mass called excitons – explains Barbara Pietka from the Polaritone Laboratory at the Faculty of Physics of the University of Warsaw. This strong interaction is possible when the photons and excitons are trapped together in the so-called optical microcavities, forcing repeated energy exchange between them. This type of synergy created in the microcavity between a photon and an exciton is so persistent that physicists call it a quasiparticle and refer to it as an excision polariton (or polariton for short).

Polaritons have unique properties, notably they can phase transition into a Bose-Einstein condensate under the right conditions. In such a state, the previously independent multiple polaritons become indistinguishable. – Based on our last experiment, we were the first to notice that polaritons, when excited by laser pulses, emit pulses of light in a way that mimics the spiking of biological neurons – describes Magdalena Furman, Ph.D. Student conducting research at the Polariton Laboratory, Faculty of Physics, University of Warsaw. This effect is directly related to the phenomenon of Bose-Einstein condensation, which either inhibits or enhances the emission of pulses.

Andrzej Opala from the Institute of Physics of the Polish Academy of Sciences, who together with Michal Matuszewski developed the theoretical basis to combine research on polaritons with the LIF model of a neuron (Leaky Integrate and Fire model), adds that the group is now working on solving the problem of scalability, that is, connecting many neurons into a network. – We propose to use a new computational paradigm based on encoding information with impulses that trigger a signal only when it arrives at the neuron at the right time one after the other – explains the researcher. Currently, neural networks use layers of interconnected neurons that trigger impulses based on the importance assigned to each connection (in the mathematical description we refer to “weights”). In contrast to this type of solution, in the optical neural network developed by researchers from Poland and described in the journal Laser and Photonics Review, the neurons are triggered (ie active) in response to an impulse train that may be present. different intensity and different time intervals. As with biological neurons, which are stimulated by electrical impulses, there is a certain threshold above which this train of impulses reaching the neuron triggers a signal that is passed on. Polaritons make it possible to mimic a biological system, because only excitation with the appropriate number of photons leads to the formation of Bose-Einstein condensate above a certain threshold and subsequent emission of a short flash in the picosecond range, which is a signal for the next neuron.

Importantly, the sample, used by scientists to capture photons and observe exciton-polaritons condensate, was synthesized on-site at the Faculty of Physics at the University of Warsaw in the group of Wojciech Pacuski. The scientists arranged the atoms of different types of semiconductor crystals layer by layer by molecular beam epitaxy to create a prototype photonic neuron. A temperature of 4K (the liquid helium) was required to reach the Bose-Einstein condensate state. – Our further goal is to transfer the experiment from cryogenic conditions to room temperature – says Jacek Szczytko from the Faculty of Physics at the University of Warsaw. – New materials that make it possible to obtain Bose-Einstein condensates even at high temperatures need to be researched. In order for photonic neurons to network, they must be able to transmit signals to one another. Ideally, the direction of transmission, i.e. the circuit diagram, could be easily changed if necessary.

– Scientists still face new challenges in the study of neuromorphic systems. Our new idea of ​​emulating the spiking of biological neurons in the optical domain can be used to create a network and then a neuromorphic system in which information is sent orders of magnitude faster and more energy-efficiently compared to existing solutions – concludes Krzysztof Tyszka.

An international team of scientists conducted research supported by National Science Center (Grants 2020/37B/ST3/01657, 2020/04/X/ST7 01379, 2020/36/T/ST3/00417), Center for Atomic, Molecular and Optical Physics and the EU program FET-Open Horizon 2020 grant “TopoLight” (964770) were supported.

Physics and astronomy at the University of Warsaw appeared in 1816 as part of the then Philosophy Faculty. In 1825 the astronomical observatory was founded. Currently, the Faculty of Physics at the University of Warsaw consists of the following institutes: Experimental Physics, Theoretical Physics, Geophysics, the Department of Mathematical Methods in Physics and the Astronomical Observatory. The research covers almost all areas of modern physics, on scales from quantum theory to cosmology. The Faculty’s research and teaching staff consists of over 200 academic teachers, 81 of whom are professors. Around 1,000 students and over 170 doctoral students study at the Faculty of Physics at the University of Warsaw.

SCIENTIFIC PUBLICATION:

Tyszka K, Furman M, Mirek R, Krol M, Opala A, Seredynski B, Suffczynski J, Pacuski W, Matuszewski M, Szczytko J, Pietka B Leaky integrate-and-fire mechanism in exciton-polariton condensates for photonic spiking neurons

Laser & Photonics Reviews 2022, 2100660
https://doi.org/10.1002/lpor.202100660

CONTACT:

Krzysztof Tyszka
Faculty of Physics, University of Warsaw
Email: [email protected]
Telephone: +48 22 55 32 749

RELATED WEBSITES WWW:

http://polariton.fuw.edu.pl/
Polariton Group website

https://www.fuw.edu.pl/faculty-of-physics-home.html
Website of the Faculty of Physics, University of Warsaw

https://www.fuw.edu.pl/press-releases.html
Press Service of the Faculty of Physics of the University of Warsaw

GRAPHIC MATERIALS:

FUW221024b_fot02
https://www.fuw.edu.pl/tl_files/press/images/2022/FUW221024b_fot02.png
Optical microcavity as a pulsating neuron (Visualization: Mateusz Krol, Source: Faculty of Physics, University of Warsaw)