High-efficiency reinforcement learning with hybrid architecture photonic integrated circuit.

Reinforcement learning (RL) stands as one of the three fundamental paradigms within machine learning and has made a substantial leap to build general-purpose learning systems. However, using traditional electrical computers to simulate agent-environment interactions in RL models consumes tremendous computing resources, posing a significant challenge to the efficiency of RL. Here, we propose a universal framework that utilizes a photonic integrated circuit (PIC) to simulate the interactions in RL for improving the algorithm efficiency. High parallelism and precision on-chip optical interaction calculations are implemented with the assistance of link calibration in the hybrid architecture PIC. By introducing similarity information into the reward function of the RL model, PIC-RL successfully accomplishes perovskite materials synthesis task within a 3472-dimensional state space, resulting in a notable 56% improvement in efficiency. Our results validate the effectiveness of simulating RL algorithm interactions on the PIC platform, highlighting its potential to boost computing power in large-scale and sophisticated RL tasks.

Observation of a Photonic Orbital Gauge Field.

Gauge field is widely studied in natural and artificial materials. With an effective magnetic field for uncharged particles, many intriguing phenomena are observed in several systems like photonic Floquet topological insulator. However, previous researches about the gauge field mostly focus on limited dimensions such as the Dirac spinor in graphene materials. Here, an orbital gauge field based on photonic triangular lattices is first proposed and experimentally observed. Disclination defects with Frank angle Ω created on such lattices breaks the original lattice symmetry and generates purely geometric gauge field operating on orbital basis functions. Interestingly, it is found that bound states near zero energy with the orbital angular momentum (OAM) l = 2 are intensively confined at the disclination as gradually expanding Ω. Moreover, the introduction of a vector potential field breaks the time-reversal symmetry of the orbital gauge field, experimentally manifested by the chiral transmission of light on helical waveguides. The orbital gauge field further suggests fantastic applications of manipulating the vortex light in photonic integrated devices.

Experimental 61-partite entanglement on a three-dimensional photonic chip.

Multipartite entanglements are essential resources for proceeding tasks in quantum information science and technology. However, generating and verifying them present significant challenges, such as the stringent requirements for manipulations and the need for a huge number of building-blocks as the systems scale up. Here, we propose and experimentally demonstrate the heralded multipartite entanglements on a three-dimensional photonic chip. Integrated photonics provide a physically scalable way to achieve an extensive and adjustable architecture. Through sophisticated Hamiltonian engineering, we are able to control the coherent evolution of shared single photon in the multiple spatial modes, dynamically tuning the induced high-order W-states of different orders in a single photonic chip. Using an effective witness, we successfully observe and verify 61-partite quantum entanglements in a 121-site photonic lattice. Our results, together with the single-site-addressable platform, offer new insights into the accessible size of quantum entanglements and may facilitate the developments of large-scale quantum information processing applications.

Classification of quantum correlation using deep learning.

Quantum correlation, as an intrinsic property of quantum mechanics, has been widely employed to test the fundamental physical principles and explore the quantum-enhanced technologies. However, such correlation would be drowned and even destroyed in the conditions of high levels of loss and noise, which drops into the classical realm and renders quantum advantage ineffective. Especially in low light conditions, conventional linear classifiers are unable to extract and distinguish quantum and classical correlations with high accuracy. Here we experimentally demonstrate the classification of quantum correlation using deep learning to meet the challenge in the quantum imaging scheme. We design the convolutional neural network to learn and classify the correlated photons efficiently with only 0.1 signal photons per pixel. We show that decreasing signal intensity further weakens the correlation and makes an accurate linear classification impossible, while the deep learning method has a strong robustness of such task with the accuracy of 99.99%. These results open up a new perspective to optimize the quantum correlation in low light conditions, representing a step towards diverse applications in quantum-enhanced measurement scenarios, such as super-resolution microscope, quantum illumination, etc.

Experimental Boson Sampling Enabling Cryptographic One-Way Function.

Boson sampling is a computational problem, which is commonly believed to be a representative paradigm for attaining the milestone of quantum advantage. So far, massive efforts have been made to the experimental large-scale boson sampling for demonstrating this milestone, while further applications of the machines remain a largely unexplored area. Here, we investigate experimentally the efficiency and security of a cryptographic one-way function that relies on coarse-grained boson sampling, in the framework of a photonic boson-sampling machine fabricated by a femtosecond laser direct writing technique. Our findings demonstrate that the implementation of the function requires moderate sample sizes, which can be over 4 orders of magnitude smaller than the ones predicted by the Chernoff bound; whereas for numbers of photons n≥3 and bins d∼poly(m,n), the same output of the function cannot be generated by nonboson samplers. Our Letter is the first experimental study that deals with the potential applications of boson sampling in the field of cryptography and paves the way toward additional studies in this direction.

Enhancing photon generation rate with broadband room-temperature quantum memory.

Photons with high generation rate is one of the essential resources for quantum communication, quantum computing and quantum metrology. Due to the naturally memory-built-in feature, the memory-based photon source is a promising route towards large-scale quantum information processing. However, such photon sources are mostly implemented in extremely low-temperature ensembles or isolated systems, limiting its physical scalability. Here we realize a single-photon source based on a far off-resonance Duan-Lukin-Cirac-Zoller quantum memory at broadband and room-temperature regime. By harnessing high-speed feedback control and repeat-until-success write process, the photon generation rate obtains considerable enhancement up to tenfold. Such a memory-enhanced single-photon source, based on the broadband room-temperature quantum memory, suggests a promising way for establishing large-scale quantum memory-enabled network at ambient condition.

Direct Observation of Dynamically Localized Quantum Optical States.

Quantum-correlated biphoton states play an important role in quantum communication and processing, especially considering the recent advances in integrated photonics. However, it remains a challenge to flexibly transport quantum states on a chip, when dealing with large-scale sophisticated photonic designs. The equivalence between certain aspects of quantum optics and solid-state physics makes it possible to utilize a range of powerful approaches in photonics, including topologically protected boundary states, graphene edge states, and dynamic localization. Optical dynamic localization allows efficient protection of classical signals in photonic systems by implementing an analogue of an external alternating electric field. Here, we report on the observation of dynamic localization for quantum-correlated biphotons, including both the generation and the propagation aspects. As a platform, we use sinusoidal waveguide arrays with cubic nonlinearity. We record biphoton coincidence count rates as evidence of robust generation of biphotons and demonstrate the dynamic localization features in both spatial and temporal space by analyzing the quantum correlation of biphotons at the output of the waveguide array. Experimental results demonstrate that various dynamic modulation parameters are effective in protecting quantum states without introducing complex topologies. Our Letter opens new avenues for studying complex physical processes using photonic chips and provides an alternative mechanism of protecting communication channels and nonclassical quantum sources in large-scale integrated quantum optics.

Reducing circuit complexity in optical quantum computation using 3D architectures.

Integrated photonic architectures based on optical waveguides are one of the leading candidates for the future realisation of large-scale quantum computation. One of the central challenges in realising this goal is simultaneously minimising loss whilst maximising interferometric visibility within waveguide circuits. One approach is to reduce circuit complexity and depth. A major constraint in most planar waveguide systems is that beamsplitter transformations between distant optical modes require numerous intermediate SWAP operations to couple them into nearest neighbour proximity, each of which introduces loss and scattering. Here, we propose a 3D architecture which can significantly mitigate this problem by geometrically bypassing trivial intermediate operations. We demonstrate the viability of this concept by considering a worst-case 2D scenario, where we interfere the two most distant optical modes in a planar structure. Using femtosecond laser direct-writing technology we experimentally construct a 2D architecture to implement Hong-Ou-Mandel interference between its most distant modes, and a 3D one with corresponding physical dimensions, demonstrating significant improvement in both fidelity and efficiency in the latter case. In addition to improving fidelity and efficiency of individual non-adjacent beamsplitter operations, this approach provides an avenue for reducing the optical depth of circuits comprising complex arrays of beamsplitter operations.

Inhibition and Reconstruction of Zener Tunneling in Photonic Honeycomb Lattices.

Quantum coherence is the central element of particle states, and it characterizes the overall performance of various quantum materials. Bloch oscillation is a fundamental coherent behavior of particles under a static potential, which can be easily destroyed by Zener tunneling in multiband 2D lattice materials. The control of Zener tunneling therefore plays the key role in quantum engineering for complicated physical systems. Here, the inhibition and reconstruction of Zener tunneling in photonic honeycomb lattices are experimentally demonstrated.  Deformed honeycomb lattices are integrated and an effective static potential is realized on the 2D lattice materials. Zener tunneling disappears in stretch-type lattices and wave packets stay in the dispersionless upper energy band. On the contrary, Zener tunneling is greatly enhanced in compression-type lattices and wave packets exhibit directional oscillations without branches, which manifest the preserved coherence of the wave packets. The results demonstrate the protection of photonic coherence by structurally controlling the Zener tunneling, representing a step toward flexible quantum engineering for large-scale artificial quantum materials.

Generating Haar-Uniform Randomness Using Stochastic Quantum Walks on a Photonic Chip.

As random operations for quantum systems are intensively used in various quantum information tasks, a trustworthy measure of the randomness in quantum operations is highly demanded. The Haar measure of randomness is a useful tool with wide applications, such as boson sampling. Recently, a theoretical protocol was proposed to combine quantum control theory and driven stochastic quantum walks to generate Haar-uniform random operations. This opens up a promising route to converting classical randomness to quantum randomness. Here, we implement a two-dimensional stochastic quantum walk on the integrated photonic chip and demonstrate that the average of all distribution profiles converges to the even distribution when the evolution length increases, suggesting the 1-pad Haar-uniform randomness. We further show that our two-dimensional array outperforms the one-dimensional array of the same number of waveguide for the speed of convergence. Our Letter demonstrates a scalable and robust way to generate Haar-uniform randomness that can provide useful building blocks to boost future quantum information techniques.

Experimentally Detecting Quantized Zak Phases without Chiral Symmetry in Photonic Lattices.

Symmetries play a major role in identifying topological phases of matter and in establishing a direct connection between protected edge states and topological bulk invariants via the bulk-boundary correspondence. One-dimensional lattices are deemed to be protected by chiral symmetry, exhibiting quantized Zak phases and protected edge states, but not for all cases. Here, we experimentally realize an extended Su-Schrieffer-Heeger model with broken chiral symmetry by engineering one-dimensional zigzag photonic lattices, where the long-range hopping breaks chiral symmetry but ensures the existence of inversion symmetry. By the averaged mean displacement method, we detect topological invariants directly in the bulk through the continuous-time quantum walk of photons. Our results demonstrate that inversion symmetry protects the quantized Zak phase but edge states can disappear in the topological nontrivial phase, thus breaking the conventional bulk-boundary correspondence. Our photonic lattice provides a useful platform to study the interplay among topological phases, symmetries, and the bulk-boundary correspondence.

Thresholded single-photon underwater imaging and detection.

Optical underwater target imaging and detection have been a tough but significant challenge in deep-sea exploration. Distant reflected signals drown in various underwater noises due to strong absorption and scattering, resulting in degraded image contrast and reduced detection range. Single-photon feature operating at the fundamental limit of the classical electromagnetic waves can broaden the realm of quantum technologies. Here we experimentally demonstrate a thresholded single-photon imaging and detection scheme to extract photon signals from the noisy underwater environment. We reconstruct the images obtained in a high-loss underwater environment by using photon-limited computational algorithms. Furthermore, we achieve a capability of underwater detection down to 0.8 photons per pulse at Jerlov type III water up to 50 meters, which is equivalent to more than 9 attenuation lengths. The results break the limits of classical underwater imaging and detection and may lead to many quantum-enhanced applications, like air-to-sea target tracking and deep-sea optical exploration.

Quantum superposition demonstrated higher-order topological bound states in the continuum.

Higher-order topological insulators, as newly found non-trivial materials and structures, possess topological phases beyond the conventional bulk-boundary correspondence. In previous studies, in-gap boundary states such as the corner states were regarded as conclusive evidence for the emergence of higher-order topological insulators. Here, we present an experimental observation of a photonic higher-order topological insulator with corner states embedded into the bulk spectrum, denoted as the higher-order topological bound states in the continuum. Especially, we propose and experimentally demonstrate a new way to identify topological corner states by exciting them separately from the bulk states with photonic quantum superposition states. Our results extend the topological bound states in the continuum into higher-order cases, providing an unprecedented mechanism to achieve robust and localized states in a bulk spectrum. More importantly, our experiments exhibit the advantage of using the time evolution of quantum superposition states to identify topological corner modes, which may shed light on future exploration between quantum dynamics and higher-order topological photonics.

Symmetry-Induced Error Filtering in a Photonic Lieb Lattice.

Quantum computation promises intrinsically parallel information processing capacity by harnessing the superposition and entanglement of quantum states. However, it is still challenging to realize universal quantum computation due that the reliability and scalability are limited by unavoidable noises on qubits. Nontrivial topological properties like quantum Hall phases are found capable of offering protection, but require stringent conditions of topological band gaps and broken time-reversal symmetry. Here, we propose and experimentally demonstrate a symmetry-induced error filtering scheme, showing a more general role of geometry in protection mechanism and applications. We encode qubits in a superposition of two spatial modes on a photonic Lieb lattice. The geometric symmetry endows the system with topological properties featuring a flat band touching, leading to distinctive transmission behaviors of π-phase qubits and 0-phase qubits. The geometry exhibits a significant effect on filtering phase errors, which also enables it to monitor phase deviations in real time. The symmetry-induced error filtering can be a key element for encoding and protecting quantum states, suggesting an emerging field of symmetry-protected universal quantum computation and noisy intermediate-scale quantum technologies.

Topologically protecting quantum resources with sawtooth lattices.

The inevitable noise and decoherence in the quantum circuit hinder its scalable development, so quantum error correction and quantumness protection for multiple controllable qubits system are necessary. The flatband in the dispersion relation, based on its inherent locality and high degenerate energy band structure, shows non-diffractive transport properties in the line spectrum and has the potential possibility to protect quantum resources in special lattices. The pioneer work has proved that the topologically boundary state is robust to protect the quantumness from disorder and perturbation, which inspires that quantumness can be protected anywhere in a periodic structure, including the boundary state and bulk state. Here, we show the topological protection of quantum resources with different state combinations in a sawtooth lattice. Photons can be localized at any degenerate eigenmode, and the localized effect is determined by only one parameter, without additional modulations. We show a high violation of Cauchy-Schwarz inequality up to 35 standard deviations by measuring cross correlation and auto-correlation of correlated photons. We verify that the topological protection is robust to different wavelengths of correlated photons. Our results suggest an alternative way of exploring topological protection in flatband and bulk state, demonstrating the powerful ability of topological photonics to protect quantum resources.

Waveguide-coupled superconducting nanowire single-photon detectors based on femtosecond laser direct writing.

The implementation of quantum information technologies requires the development of integrated quantum chips. Femtosecond laser direct writing (FLDW) waveguides and superconducting nanowire single-photon detectors (SNSPDs) have been widely applied in integrated quantum photonic circuits. In this work, a novel FLDW waveguide-coupled SNSPD was designed and realized by integrating FLDW waveguides and conventional SNSPDs together. Through a COMSOL simulation, a waveguide end face-nanowire optical coupling structure was designed and verified. The simulation results showed that the FLDW waveguide-coupled SNSPD device, which had a target wavelength of 780 nm, can achieve 87% optical absorption. Then the preparation process of the FLDW waveguide-coupled SNSPD device was developed, and the fabricated device achieved a system detection efficiency of 1.7% at 10 Hz dark count rate. Overall, this method provides a feasible single-photon detector solution for future on-chip integrated quantum photonic experiments and applications.

TensorFlow solver for quantum PageRank in large-scale networks.

Google PageRank is a prevalent algorithm for ranking the significance of nodes or websites in a network, and a recent quantum counterpart for PageRank algorithm has been raised to suggest a higher accuracy of ranking comparing to Google PageRank. The quantum PageRank algorithm is essentially based on quantum stochastic walks and can be expressed using Lindblad master equation, which, however, needs to solve the Kronecker products of an O(N4) dimension and requires severely large memory and time when the number of nodes N in a network increases above 150. Here, we present an efficient solver for quantum PageRank by using the Runge-Kutta method to reduce the matrix dimension to O(N2) and employing TensorFlow to conduct GPU parallel computing. We demonstrate its performance in solving quantum stochastic walks on Erdös-Rényi graphs using an RTX 2060 GPU. The test on the graph of 6000 nodes requires a memory of 5.5 GB and time of 223 s, and that on the graph of 1000 nodes requires 226 MB and 3.6 s. Compared with QSWalk, a currently prevalent Mathematica solver, our solver for the same graph of 1000 nodes reduces the required memory and time to only 0.2% and 0.05%. We apply the solver to quantum PageRank for the USA major airline network with up to 922 nodes, and to quantum stochastic walk on a glued tree of 2186 nodes. This efficient solver for large-scale quantum PageRank and quantum stochastic walks would greatly facilitate studies of quantum information in real-life applications.

Real-space observation of topological invariants in 2D photonic systems.

Topological materials are capable of inherently robust transport and propagation of physical fields against disorder and perturbations, holding the promise of revolutionary technologies in a wide spectrum. Higher-order topological insulators are recently predicted as topological phases beyond the standard bulk-edge correspondence principle, however, their topological invariants have been proven very challenging to observe, even not possible yet by indirect ways. Here, we demonstrate theoretically and experimentally that the topological invariants in two-dimensional systems can be directly revealed in real space by measuring single-photon bulk dynamics. By freely writing photonic lattices with femtosecond laser, we construct and identify the predicted second-order topological insulators, as well as first-order topological insulators with fractional topological winding number. Furthermore, we show that the accumulation and statistics on individual single-particle registrations can eventually lead to the same results of light waves, despite the fact that the development of topological physics was originally based on wave theories, sharing the same spirit of wave-particle nature in quantum mechanics. Our results offer a direct fashion of observing topological phases in two-dimensional systems and may inspire topologically protected artificial devices in high-order topology, high-dimension and quantum regime.

Integrated Quantum-Walk Structure and NAND Tree on a Photonic Chip.

In the age of the post-Moore era, the next-generation computing model would be a hybrid architecture consisting of different physical components, such as photonic chips. In 2008, it was proposed that the solving of the NAND-tree problem can be sped up by quantum walk. This scheme is groundbreaking due to the universality of the NAND gate. However, experimental demonstration has not been achieved so far, mostly due to the challenge in preparing the propagating initial state. Here we propose an alternative solution by including a structure called a "quantum slide," where a propagating Gaussian wave packet can be generated deterministically along a properly engineered chain. In our experimental demonstration, the optical NAND tree is capable of solving computational problems with a total of four input bits, based on the femtosecond laser 3D direct-writing technique on a photonic chip. These results remove one main roadblock to photonic NAND-tree computation, and the construction of a quantum slide may find other interesting applications in quantum information and quantum optics.

Single-photon characterization by two-photon spectral interferometry.

Single-photon sources are a fundamental resource in quantum optics and quantum information science. Photons with differing spectral and temporal shapes do not interfere well and inhibit the performance of quantum applications such as linear optics quantum computing, boson sampling, and quantum networks. Indistinguishability and purity of photons emitted from different sources are crucial properties for many quantum applications. The ability to determine the state of single-photon sources therefore provides a means to assess their quality, compare different sources, and provide feedback for source tuning. Here, we propose and demonstrate a single-configuration experimental method enabling complete characterization of the spectral-temporal state of a pulsed single-photon source having both pure and mixed states. The method involves interference of the unknown single-photon source with a reference at a balanced beam splitter followed by frequency-resolved coincidence detection at the outputs. Fourier analysis of the joint-spectral two-photon interference pattern reveals the density matrix of the single-photon source in the frequency basis. We present an experimental realization of this method for pure and mixed state pulsed single-photon sources.