Researcher Jelmer Renema of the University of Twente is a member of a Chinese research team that developed a new photonics-based quantum computer showing quantum advantage – it can solve problems a supercomputer can’t. Compared to its predecessor, the new quantum computer shows lower losses and can measure more photons, leading to a new record speed of a trillion squared compared to brute force supercomputing. Renema contributed to this by introducing better theoretical understanding. The results of the ‘Jiuzhang 2.0’ computer are published in Physical Review Letters.
A quantum computer that outperforms the conventional super computer is no longer a theoretical exercise. It can be done using superconducting ‘qubits’, like Google’s Sycamore system does. Another emerging technology for building quantum computers is photonics. Using photons, light particles, instead of qubits, has the advantage of higher stability. And a photonic quantum computer can, for the major part, run at room temperature whereas qubits only function at very low temperatures. The photonic quantum computer that is now presented, can solve a complex problem that a supercomputer can’t within a realistic time frame.
Pin ball machine
The calculation is called Gaussian Boson Sampling. At one end of the computer, there are light sources that can emit a single photon at a time, at the output, there are 144 detectors that can each detect one photon. In between is a complex structure of mirrors, lenses and chips that manipulate light. This chip, for example has splitters that cause photons to move in two directions at the same time. These light paths connect to others, so photons interfere. In the end, the photons moving out of the system, are detected one at a time. The distribution of the photons, the actual outcome of the calculation, is shown after repeating the exercise for a large number of times. If a photon would be represented by a ball, the experiment is a pin ball machine with 100 balls, in which a ball falls in any of 144 bins. It may be hard to imagine the practical relevance of the calculation, but according to the team, it already comes close to solving problems in chemistry and in quantum machine learning.
The paths that photons travel between the light sources and detectors, with all quantum effects taking place, can hardly be ‘followed’ using brute calculation force. The paper mentions a ‘sampling rate’ that is 1024 higher than that of a state-of-art supercomputer, for this. According to Renema, this result is not only underlining the potential of photonics in quantum computing, it is also ahead of any of the road maps in quantum computing so far.
Jelmer Renema, researcher within the Adaptive Quantum Optics Group (MESA+/QUANT), is also one of the founders and Chief Technology Officer (CTO) at QuiX Quantum. This company develops light conducting chips that can be used as quantum processors. ‘Quantum daily’ recently mentioned him in a top 51 of CTO’s to keep an eye on in the world of quantum. QuiX is also in the ‘Silicon 100: Startups worth watching in 2021’.
The paper ‘Phase-programmable Gaussian Boson Samping using stimulated squeezed light’ by Han-Sen Zhon, Yu-Hoan Doeng, Jian Qin, Hui Wang, Min-Cheng Chen, Li-Chao Peng, Yi-Han Luo, Dian Wu, Si-Qu
iu Gong, Hao Su, Yi Hu, Peng Hu, Xiao-Yan Yang, Wei-Jun Zhang, Hao Li, Yuxuan Li, Xiao Jiang, Li Gan, Huangwen Yang, Lixing You, Zhen Wang, Li Li, Nai-Le- Liu, Jelmer Renema, Chao-Yang Lu en Jian-Wei Pan, is in Physical Review Letters. It is also highlighted in 'Physics': 'Quantum leap for quantum primacy'.