Shanghai Jiaotong University Demonstrates Quantum Superiority of Amnesic Resistor Boson Sampling

Recently, the latest research results of Xianmin Jin's team at Shanghai Jiao Tong University, entitled "Quantum advantage with membosonsampling"¹ were published in Chip, with Jun Gao, Xiaowei Wang, and Wenhao Zhou as co-first authors and Xianmin Jin as corresponding author. Chip is the only comprehensive international journal focusing on chip research in the world.

The team has proposed a new bosonic sampling scheme inspired by the "memristor" mechanism, which allows quantum interference effects to occur between time periods through a circular structure, thereby increasing the overall computational complexity. The team succeeded in extracting 56 photon events in 750,000 patterns, thus demonstrating an integrated and cost-effective scheme in photonic systems, leading to "quantum superiority". At the same time, the results provide a controllable and scalable platform for quantum simulations in very large Hilbert spaces.

Fig. 1 Schematic diagram of memristor boson sampling (Membosonsampling). Each time block contains n photons and m modes, and grows linearly with the number of repetitions N. Through the structure of the cycle, the time blocks are coherently connected and influence each other. Quantum interference effects within the same time block and between different time blocks are experimentally recorded and analyzed separately and independently.

Since Google's announcement in 2019 of "quantum superiority"² through the superconducting quantum processor Sycamore, quantum computing has entered the era of Noisy Intermediate-Scale Quantum (NISQ)³, despite some controversy. Boson sampling⁴ has long been considered as another candidate for achieving "quantum superiority". Optical platforms are very friendly for implementing Boson sampling experiments due to their long photon coherence time and good robustness. The Ninth Chapter optical quantum computer⁵ has already achieved "quantum superiority" by Gaussian bosonic sampling. However, it is still a challenge to build a controllable, practical, and economical quantum system that can be mapped to practical applications while demonstrating computational challenges in its own right.

Fig. 2 Schematic diagram of the memristor bosonic sampler. The multi-photon SPDC light source's are collected by fiber coupler and injected into the integrated photonic chip, and by expanding the time block, the number of photons and modes can be expanded by a factor of N. The circular structure ensures that quantum interference effects can occur between different time blocks. The photon outputs of different layers are detected by the single-photon detector array and recorded by the time-of-flight module, and all the time information can be recorded and analyzed independently.

The research team proposed a new bosonic sampling scheme inspired by the "memristor" mechanism, called "memristor bosonic sampling" (membosonsampling). The circular structure allows the quantum interference effects within the same time block and between different time blocks to be recorded and analyzed independently. Each time block contains n photons and m modes, and the number of photons and modes of the experiment grows linearly with the number of repetitions N through multiple time reuse. The scheme successfully fuses random scattershot boson sampling (scattershot bosonsampling) with temporal degrees of freedom, increasing the overall computational complexity. Moreover, in principle, the scale of the problem can be extended to infinity.

Figure 3 Example of 56 photon events on a chip. The gray bars represent the time layers where the photon events are located, and the red squares represent the photon interference between different time layers.

The research team verified the scheme by the interference of multiple SPDC light sources in a cyclic structured photonic chip. The multiphoton SPDC light sources are collected by a fiber coupler and injected into the integrated photonic chip, and the number of photons and modes can be expanded by a factor of N by expanding the number of time cycles. The self-cycling structure of the photonic chip ensures that quantum interference effects can also occur between different time blocks. The time-of-flight module records the temporal information of all photon events in the huge Hilbert space and is used to extract the outgoing photon probability distribution. The results of the photon probability distribution of the temporally random scattering output are successfully distinguished from the classical resolvable sampler. By adding blocks of time, the problem is extended to 200,000 photons scattered in 750,000 modes, and the on-chip 56 photon events are successfully restored, reaching the category of "quantum superiority".

This work demonstrates an integrated and cost-effective new path toward "quantum superiority" in photonic systems and provides a new platform that can be controlled and extended for quantum simulations in very large Hilbert spaces.

About Chip

Chip is the world's only comprehensive international journal focusing on chip-based research, and has been selected as one of the "Three High-Quality Papers It is one of the journals encouraged by the Ministry of Science and Technology to publish "three types of high-quality papers".

Chip is published by Shanghai Jiao Tong University in cooperation with Elsevier Group, and cooperates with many famous academic organizations at home and abroad to provide a high-quality communication platform for academic conferences.

Chip adheres to the founding philosophy: All About Chip, focusing on chips and being inclusive, aiming to publish cutting-edge breakthroughs in various scientific research fields related to chips and help the future development of chip technology. So far, Chip has brought together 68 world-renowned experts and scholars from 13 countries in its editorial board, including many Chinese and foreign academicians and life fellows of IEEE, ACM, Optica and other well-known international societies.

Links to papers：

https://www.sciencedirect.com/science/article/pii/S2709472322000053

Reference：

[1]Gao, J. et al. Quantum advantage with membosonsampling.Chip 1, 100007 (2022).

[2]Arute, F. et al. Quantum supremacy using a programmable superconducting processor.Nature 574, 505-510 (2019).

[3]Preskill, J. Quantum Computing in the NISQ era and beyond.Quantum 2, 79 (2018).

[4]Aaronson, S. & Arkhipov, A. The computational complexity of linear optics.Theory Comput. 9, 143-252.

[5]Zhong, H.-S. et al. Quantum computational advantage using photons.Science 370, 1460-1463.