A5004536619- Quantum optics and atomic interactions
- Quantum Information and Cryptography
- Cold Atom Physics and Bose-Einstein Condensates
- Optical Network Technologies
- Neural Networks and Reservoir Computing
- Photonic and Optical Devices
- Atomic and Subatomic Physics Research
- Quantum Mechanics and Applications
- Advanced Fluorescence Microscopy Techniques
- Orbital Angular Momentum in Optics
- Advanced Fiber Laser Technologies
- Semiconductor Lasers and Optical Devices
- Advanced Memory and Neural Computing
- Mechanical and Optical Resonators
- Neural Networks and Applications
- Advanced Thermodynamics and Statistical Mechanics
- Vehicle License Plate Recognition
- Luminescence Properties of Advanced Materials
- Random lasers and scattering media
- Semiconductor Quantum Structures and Devices
- Image and Object Detection Techniques
- Spectroscopy Techniques in Biomedical and Chemical Research
- Nonlinear Photonic Systems
- Quantum Mechanics and Non-Hermitian Physics
- Optical Coherence Tomography Applications
University of Oxford
2019-2025
Shaoxing University
2021-2024
Oxfam
2022
University of Electronic Science and Technology of China
2020-2021
Science Oxford
2021
University of Calgary
2021
Hong Kong University of Science and Technology
2013-2019
University of Hong Kong
2014-2019
University of Bergen
2015
Zhejiang University
2013
Knot detection is a challenging problem for the wood industry. Traditional methodologies depend heavily on features selected manually and therefore were not always accurate due to variety of knot appearances. This paper proposes an automated framework addressing aforementioned by using state-of-the-art YOLO-v5 (the fifth version You Only Look Once) detector. The surface knots learned extracted adaptively, then defects identified accurately even though vary in terms color texture. proposed...
We propose a practical scheme for end-to-end optical backpropagation in neural networks. Using saturable absorption the nonlinear units, we find that backward-propagating gradients required to train network can be approximated surprisingly simple pump-probe requires only passive elements. Simulations show that, with readily obtainable depths, our approach achieve equivalent performance state-of-the-art computational networks on image classification benchmarks, even deep multiple sequential...
We report an experiment demonstrating the generation of directional thermal radiation with a spectral brightness that is about 9 times greater than ambient pumping reservoir. The based on recent proposal for nontraditional quantum heat engine and uses cold Rb atoms, electromagnetically induced transparency, photon correlation spectroscopy [Phys. Rev. A 94, 053859 (2016)PLRAAN2469-992610.1103/PhysRevA.94.053859].
The generation of nonclassical photon pairs with a long coherence time is key for applications that range from fundamental to quantum communication and metrology. Spontaneous four-wave mixing electromagnetically induced transparency has been demonstrated as one the most efficient methods; however, narrowing bandwidth producing temporal length beyond 1 μs remains technical challenge due noise considerations need cold atoms high optical depth (OD). In this work, we demonstrate narrowband...
Optics is a promising platform in which to help realize the next generation of fast, parallel, and energy-efficient computation. We demonstrate reconfigurable free-space optical multiplier that capable over 3000 computations using spatial light modulators with pixel resolution only 340×340. This enables vector-matrix multiplication parallel vector-vector vector size up 56. Our design is, best our knowledge, first simultaneously support implementation reconfigurable, large-sized, real-valued...
Optical neural networks are emerging as a promising type of machine learning hardware capable energy-efficient, parallel computation. Today’s optical mainly developed to perform inference after in silico training on digital simulators. However, various physical imperfections that cannot be accurately modeled may lead the notorious “reality gap” between simulator and system. To address this challenge, we demonstrate hybrid where weight matrix is trained with neuron activation functions...
Nanophosphors of (YAG/Ce) were synthesized with a novel salted sol-gel (SSG) method in which water solution inorganic salt, yttrium nitrate [, YNO], was used traditional metal alkoxide precursor, aluminum sec-butoxide ASB], synthesis. With the SSG method, YAG single phase could be obtained by sintering dry gel and mixture for at . The particle size range from Luminescence properties samples different doping concentrations studied. peak intensity luminescence found 4% concentration. Red shift...
We demonstrate a technique for shaping the temporal wave function of biphotons generated from spatially modulated spontaneous four-wave mixing in cold atoms. show that spatial profile pump field can be mapped onto biphoton group delay regime. The laser beam is shaped by using light modulator. This spatial-to-temporal mapping enables generation narrow-band with controllable waveforms.
We propose and demonstrate an approach to measuring the biphoton temporal wave function with polarization-dependent time-resolved two-photon interference. Through six sets of interference measurements projected onto different polarization subspaces, we can reconstruct amplitude phase functions waveform. For first time, apply this technique experimentally determine quantum states narrow-band biphotons generated from spontaneous four-wave mixing in cold atoms.
We demonstrate the generation of high-quality narrowband biphotons from a Doppler-broadened hot rubidium atomic vapor cell. Choosing double-Λ energy level scheme for optimizing both spontaneous four-wave mixing nonlinear parametric interaction and electromagnetically induced transparency (EIT), we achieve biphoton spectral brightness as high 14 000 s−1 MHz−1. Meanwhile, apply spatially tailored optical pumping beam reduction Raman noise obtain violation Cauchy-Schwarz inequality by factor 1023.
Entanglement, describing the inseparability of a quantum multiparty system, is one most intriguing features mechanics. Violation Bell inequality, for ruling out possibility local hidden variable theories, commonly used as strong witness entanglement. In previous test experiments with photonic entanglement based on two-photon coincidence measurement, photon temporal wave packets are absorbed completely by detectors. That is, coherence time much shorter than detection window. Here we...
This roadmap consolidates recent advances while exploring emerging applications, reflecting the remarkable diversity of hardware platforms, neuromorphic concepts, and implementation philosophies reported in field. It emphasizes critical role cross-disciplinary collaboration this rapidly evolving
We demonstrate the generation of narrowband biphotons with polarization-frequency-coupled entanglement from spontaneous four-wave mixing in cold atoms. The coupling between polarization and frequency is realized through a shifter linear optics. When polarization-frequency degrees freedom are decoupled, it robust to create Bell states, confirmed by quantum-state tomography two-photon temporal quantum beating. Making use transfer phase retardation entangled modes, we produce state tunable...
PT-symmetric optical systems have been widely researched in recent years and most of the researches are limited to weakly guiding condition. In this paper, we propose a method based on conventional coupled mode theory (CMT) describe coupling regardless conditions. With method, can examine unique PT characteristics design various structures.
The manipulation of low-energy beams neutral atoms and molecules via their de Broglie wavelength is a branch atom optics often referred to as matter wave optics. application areas include fundamental quantum mechanics, interferometry, the development new microscopy instrumentation. focusing waves with Fresnel zone plate was used demonstrate first helium imaging. ultimate resolution such microscope limited by width outermost zone. Because for cannot be fabricated on substrate (the would not...
Optics is an exciting route for the next generation of computing hardware machine learning, promising several orders magnitude enhancement in both computational speed and energy efficiency. However, to reach full capacity optical neural network it necessary that not only inference, but also training be implemented optically. The primary algorithm a backpropagation, which calculation performed order opposite information flow inference. While straightforward digital computer, implementation...
Manipulating polarization entanglement of paired photons is always great interest for understanding the quantum nature and exploring their applications in information processing communication. Narrowband biphotons with are especially important a network based on efficient photon–atom interactions. In most demonstrated cases, polarization-entangled states manipulated through birefringent effect. this Letter, we produce narrowband from spontaneous four-wave mixing cold atoms demonstrate new...
Deep learning has profoundly reshaped the technology landscape in numerous scientific areas and industrial sectors. This advancement is, nevertheless, confronted with severe bottlenecks digital computing. Optical neural network presents a promising solution due to ultra-high computing speed energy efficiency. In this work, we present systematic study of unitary optical (UONN) as an approach towards deep learning. Our results show that UONN can be trained high accuracy through special...
We report the demonstration of a configurable coherent quantum-memory-based beam splitter (BS) for single-photon wave packet making use laser-cooled $^{85}\mathrm{Rb}$ atoms and electromagnetically induced transparency. The is converted (stored) into collective atomic spin state later retrieved (split) two nearly opposing directions. storage time, beam-splitting ratio, relative phase are can be dynamically controlled. experimentally confirm that such BS preserves quantum particle nature...
Phosphor converted LED is one of the commonly used components in solid state luminaires. Among packaging processes involved, phosphor deposition a critical step, which controls overall optical performance LEDs. There are several methods, among disperse dispensing and conformal coating methods widely used. In these two materials applied directly on top chip. The heated up by chip during operation. behavior highly depends their temperature. emission efficiency decreases as temperature...
Optical neural networks are often trained “in-silico” on digital simulators, but physical imperfections that cannot be modelled may lead to a “reality gap” between the simulator and system. In this work we present hybrid training, where weight matrix is by computing neuron values optically using actual network.
We report the demonstration of a mirrorless optical parametric oscillator with tunable threshold in laser-cooled atoms four-wave mixing (FWM) using electromagnetically induced transparency. Driven by two classical laser beams, generated Stokes and anti-Stokes fields counterpropagate build up efficient intrinsic feedback through nonlinear FWM process. This does not involve any cavity or spatially distributed microstructures. observe transition photon correlation properties from biphoton...
Backpropagation through nonlinear neurons is an outstanding challenge to the field of optical neural networks and major conceptual barrier all-optical training schemes. Each neuron required exhibit a directionally dependent response propagating signals, with backwards conditioned on forward signal, which highly non-trivial implement optically. We propose practical surprisingly simple solution that uses saturable absorption provide network nonlinearity. find backward gradients train can be...