The prospect of simulating the quantum computing power, the Shanghai Jiaotong University is amazing this time.

Introduction

On May 11th, Science Advances published the latest research results of Jin Xianmin's research team of Shanghai Jiaotong University on the topic of "Experimental Two-dimensional Quantum Walk on a Photonic Chip", and reported the world's largest three-dimensional integrated light quantum chip. Demonstrates the first truly spatial two-dimensional random walk quantum computation. At the same time, this is also the first self-implemented photo quantum computing chip in China. This research progress is of great significance for advancing the research of analog quantum computers and realizing "quantum hegemony."

In recent years, news about general-purpose quantum computers has been reported in newspapers, and companies such as IBM, Google, and Intel have rushed to announce higher quantum bit counts. But the industry consensus is that even if you make dozens or more of the number of qubits, if you don't have full interconnection, the accuracy is not enough, and you can't correct it, general quantum computing can't be realized. In contrast, analog quantum computing can directly build quantum systems without the need for complex quantum error correction like general quantum computing. Once the quantum physics system that can be prepared and controlled reaches a new level, it can be directly used to explore new physics and advance the absolute computing power of far more classic computers on specific issues.

The prospect of simulating quantum computing

Analog quantum computing, compared with general quantum computing, has a more approachable physical implementation, and has obvious natural correspondence and acceleration advantages for Bose sampling, search, Hamiltonian learning, and chemical simulation. Therefore, it is another indispensable and vital area for the development of quantum information. The quantum software OpenFermion launched by Google in 2017 is dedicated to analog quantum computing.

As a powerful tool for simulating quantum computing, quantum walk in two-dimensional space can map specific computational tasks to matric coefficient matrices in quantum evolution space. Equivalent evolution systems can be prepared large enough and flexible. When it is structured, it can be used to implement various search and optimization problems in various fields such as engineering, finance, biomedicine, etc. It shows far superior performance to classical computers and has broad application prospects.

However, if you want to use quantum walking to simulate quantum computing to demonstrate the superiority of quantum algorithms, you must satisfy two points: enough walking paths, and evolutionary space that can be freely designed according to the needs of the algorithm. The previous quantum walking experiments were limited by the size limitation of the physical system that can be prepared. They can only make a principle demonstration of several small steps of evolution, and can never be freely evolved in a real spatial two-dimensional system, which is far from enough for Simulated quantum computing experiments.

Jin Xianmin's research group prepared an integrated three-dimensional waveguide chip by femtosecond laser direct writing technology. The waveguide strike represents the continuous evolution time, and the end face forms a super-large evolution space of 49×49 nodes (ie, a total of 2401 nodes), so that even single-photon injection It can also realize thousands of quantum walking paths. In experiments, quantum reaches at least one hundred walking steps, breaking through the quantum walking experimental record. At the same time, in the evolution process, the coupling of light quantum between waveguides can also be strong. Precise regulation by designing the waveguide spacing. Even precision waveguide bending, quantitative introduction loss and other regulation techniques are steadily developing. The continually sophisticated integrated waveguide chip technology enables quantum walks to move closer to practical analog quantum computing applications.

The research team observed the two-dimensional walking mode of photon by preparing PPKTP high-brightness single photon source and developing high-resolution ICCD single photon imaging technology. Experiments verify that quantum walk has a ballistic transport that is different from classical random walk in both one-dimensional and two-dimensional evolution spaces. This accelerated transmission is the basis for supporting quantum walks that can go beyond classical computers in many algorithms. The theory has pointed out that transient network characteristics are only realized in quantum walks larger than one dimension. In the past, quasi-two-dimensional quantum walk experiments could not observe network propagation characteristics due to the limited quantum evolution space. For the first time, the study observed transient network characteristics in experiments, and further verified the two-dimensional characteristics of quantum walking achieved.

Figure: Two-dimensional quantum walk evolution of single-photon from left to right: quantum walk evolution time increases

Ultra-large-scale photon computing chips are hard to come by

Jin Xianmin's team produced a three-dimensional quantum quantum computing chip with thousands of nodes through femtosecond laser direct writing technology. It is this ultra-large-scale photo-quantitative computing chip that enables true two-dimensional free-spaced quantum walking to be realized for the first time in experiments and to promote the future. More implementation of quantum algorithms.

In the twenty sets of photonic arrays in the illustrated chip, each set contains 2,401 waveguides.

These ultra-large-scale photo-quantitative computing chips are not easy to obtain. As early as 2014, Jin Xianmin gave up the opportunity to live in the UK. From Oxford University, he returned to Shanghai Jiaotong University to set up a quantum information laboratory from scratch. He set the goal of photonic chip research, from building a laboratory femtosecond. The laser direct writing platform has been continually exploring the direct writing parameters, and it took three years before and after, so that the effect of each parameter on the performance of the waveguide and how to write the waveguide performance in the demand are sufficient. At the same time, it takes two and a half years to build a high-brightness single photon source and develop a high-precision single-photon imaging technology, which makes the evolution mode of photon walking in a two-dimensional space in a chip for the first time.