A5010251203- Metamaterials and Metasurfaces Applications
- Advanced Antenna and Metasurface Technologies
- Photonic Crystals and Applications
- Plasmonic and Surface Plasmon Research
- Thermal Radiation and Cooling Technologies
- Antenna Design and Analysis
- Orbital Angular Momentum in Optics
- Photonic and Optical Devices
- Acoustic Wave Phenomena Research
- Advanced Neural Network Applications
- Optical Coatings and Gratings
- Electromagnetic Scattering and Analysis
- Advanced Optical Imaging Technologies
- Recommender Systems and Techniques
- Machine Learning and ELM
- Random lasers and scattering media
- Magneto-Optical Properties and Applications
- Electromagnetic Simulation and Numerical Methods
- Speech and Audio Processing
- Near-Field Optical Microscopy
- Domain Adaptation and Few-Shot Learning
- Microwave Engineering and Waveguides
- Advanced Materials and Mechanics
- Advanced Thermodynamics and Statistical Mechanics
- Magnetic properties of thin films
Taiyuan University of Technology
2017-2025
Harbin Institute of Technology
2022-2025
Hefei University
2020-2024
Chinese Academy of Sciences
2023-2024
Shanghai University
2024
Institute of Computing Technology
2024
Hunan Normal University
2024
Michigan State University
2023
Tongji University
2023
Qingdao University
2023
Machine-learning (ML) hardware and software system demand is burgeoning. Driven by ML applications, the number of different inference systems has exploded. Over 100 organizations are building chips, that incorporate existing models span at least three orders magnitude in power consumption five performance; they range from embedded devices to data-center solutions. Fueling a dozen or more frameworks libraries. The myriad combinations make assessing ML-system performance an...
Simultaneously manipulating multiple physical fields plays an important role in the increasingly complex integrated systems, aerospace equipment, biochemical productions, etc. For on-chip systems with high integration level, precise and efficient control of propagation electromagnetic waves heat fluxes simultaneously is particularly important. In this study, we propose a graphical designing method (i.e., thermal-electromagnetic surface transformation) based on null medium to thermal...
With the explosive growth of online information, recommender systems play a key role to alleviate such information overload. Due important application value systems, there have always been emerging works in this field. In main challenge is learn effective user/item representations from their interactions and side (if any). Recently, graph neural network (GNN) techniques widely utilized since most essentially has structure GNN superiority representation learning. This article aims provide...
Transformation optics has made a major contribution to the advancement of modern electromagnetism and related research assisted by development metamaterials. In this work, we applied concept thermodynamics using coordinate transformation time-dependent heat diffusion equation manipulate flux predefined paths. Experimentally, demonstrated transient thermal cloaking device engineered with effective materials successfully hid centimeter-sized vacuum cavity. A rescaled accounting for all...
Abstract In the modern world, ability to manipulate and control electromagnetic waves has greatly changed people's lives. Novel optical phenomena devices will lead new scientific trends techniques in future. The exploration of theories design materials for engineering attracted great attention recent years. Transformation optics (TO) provides a way with extraordinary predesigned functions such as invisibility cloaks wormholes. As development artificial media (e.g. metamaterials metasurfaces)...
Self-attention based transformer models have been dominating many computer vision tasks in the past few years. Their superb model qualities heavily depend on excessively large labeled image datasets. In order to reduce reliance datasets, reconstruction masked autoencoders are gaining popularity, which learn high quality transferable representations from unlabeled images. For same purpose, recent weakly supervised pretraining methods explore language supervision text captions accompanying...
Although previously reported terahertz absorbers can achieve high-sensitivity refractive index sensing, the resonant peak is too broad, which leads to a low figure of merit (FOM). Transmissive sensors based on bound states in continuum (BIC) high FOM, but they have some limitations sensitivity. Herein, we propose periodic triple parallel metal bars structure obtain quality, strong field, and multiple hot spots by Friedrich–Wintgen BIC. Numerical results show sensitivity FOM reach 1877...
Simultaneously guiding electromagnetic waves and heat flow at any incidence angle to smoothly bypass some electromagnetic/thermal sensitive elements is a key factor ensure efficient communication thermal protection for an on-chip system. In this study, omnidirectional electromagnetic-thermal cloak proposed. Firstly, holey metallic plate with periodic array of subwavelength apertures designed by optical surface transformation realize cloaking module signal. Secondly, two-layer ring-shaped...
Kernel Density Estimation (KDE) is an important approach to analyse spatial distribution of point features and linear over 2-D planar space. Some network-based KDE methods have been developed in recent years, which focus on estimating density events 1-D network However, the existing are not appropriate for analysing characteristics certain kind or events, such as traffic jams, queue at intersections taxi carrying passenger events. These occur distribute road space, present a continuous along...
Abstract A new theory on designing electromagnetic/optical devices is proposed, namely, an optical surface transformation (OST). One arbitrary can establish the corresponding relationship with another entirely optically optic-null medium (ONM), (i.e. electromagnetic wave propagates from one to its equivalent without any phase delay). Many novel be designed by OST help of ONM. Compared traditional Transformation Optics, our surface-reshaping have two main advantages. Firstly, design process...
Machine-learning (ML) hardware and software system demand is burgeoning. Driven by ML applications, the number of different inference systems has exploded. Over 100 organizations are building chips, that incorporate existing models span at least three orders magnitude in power consumption five performance; they range from embedded devices to data-center solutions. Fueling a dozen or more frameworks libraries. The myriad combinations make assessing ML-system performance an...
The era of AI computing brings significant challenges to traditional computer systems. As shown in Fig. 29.1.1, while the model computation requirement increases 750x every two years, we only observe a very slow-paced improvement memory system capability terms both capacity and bandwidth. There are many memory-bound applications, such as natural language processing, recommendation systems, graph analytics, neural networks, well multi-task online inference, that become dominating applications...
By introducing a columnar null-space region as the reference space, we design radome that can extend scanning angle of phased array antenna (PAA) by predetermined relationship (e.g. linear between incident and steered output be achieved). After some approximation, only need two homogeneous materials to construct proposed layer layer. This kind medium is called medium, which has been studied fabricated for realizing hyper-lenses other devices. Numerical simulations verify performance our radome.
Remote temperature control can be obtained by a long-focus thermal lens that focus heat fluxes into spot far away from the back surface of and create virtual source/sink in background material, around which field distribution remotely controlled changing parameters lens. However, due to lack negative conductivity, existing lenses have extremely short focal lengths cannot used source/sink. In this study, we theoretically propose general approach equivalently realize conductivity elaborately...
Acoustic splitters and electromagnetic can be applied in various fields (e.g., navigation interference detection). However, there is still a lack of study structures that simultaneously split acoustic beams. In this study, novel, to the best our knowledge, electromagnetic-acoustic splitter (EAS) based on copper plates proposed, which produce identical beam-splitting effects for transverse magnetic (TM)-polarized waves. Different from previous beam splitters, splitting ratio proposed passive...
Electromagnetic invisible devices usually designed by transformation optics are rather complicated in material parameters and not suitable for general applications. Recently, a topology optimized cloak based on level-set method was proposed to realize nearly perfect cloaking [Fujii et al., Appl. Phys. Lett. 102, 251106 (2013)]. In this work, we experimentally implemented idea fabricated unidirectional with relative large region made of single dielectric material. Good performance verified...
A multi-physics null medium that performs as a perfect endoscope for both electromagnetic and acoustic waves is designed by transformation optics, which opens new way to control simultaneously. Surface multi-physics, novel graphical method design devices, proposed based on the directional projecting feature of medium. Many including beam shifters, scattering reduction, imaging devices steering can be simply in surface-corresponding manner. All surface only need one homogeneous anisotropic...