Xinxin Gao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xinxin Gao has a total of 21 co-authors. Here are the top co-authors by number of shared publications:
- Qian Ma
State Key Laboratory of Millimeter Wave, Southeast University, Nanjing 210096, China
3 shared publications - Tie Jun Cui
State Key Laboratory of Millimeter Waves Southeast University Nanjing China
3 shared publications - Jian Wei You
State Key Laboratory of Millimeter Wave Southeast University Nanjing China
2 shared publications - Wen Yi Cui
State Key Laboratory of Millimeter Waves, Southeast University, Nanjing, China
2 shared publications Ze Gu
State Key Laboratory of Millimeter Waves, Southeast University, Nanjing, China
2 shared publications- Ze Gu
State Key Laboratory of Millimeter Wave, Southeast University, Nanjing, China
2 shared publications
Latest Publications
- 29 November 2025
- Published by Zhejiang University Press in Frontiers of Information Technology & Electronic Engineering
- Vol. 26 (11) , 2382-2392
Bidirectional-pump-controlled reconfigurable nonlinear spoof plasmonic waveguide
We present a dynamically reconfigurable spoof surface plasmon polariton (SSPP) waveguide capable of bidirectional switching between perfect absorption and
- 29 November 2025
- Published by Zhejiang University Press in Frontiers of Information Technology & Electronic Engineering
- Vol. 26 (11) , 2382-2392
Bidirectional-pump-controlled reconfigurable nonlinear spoof plasmonic waveguide
We present a dynamically reconfigurable spoof surface plasmon polariton (SSPP) waveguide capable of bidirectional switching between perfect absorption and
- Open Access
- 24 November 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 10332
Multilayer nonlinear diffraction neural networks with programmable and fast ReLU activation function
Optical diffractive neural networks are emerging for improving speed and energy efficiency in machine learning. However, the challenges of nonlinear activation functions (e.g., latency issues, high power consumption, and cascading complexity) impede their performance and practical deployment. Here, we propose a programmable multilayer full-space nonlinear neural network operating in the microwave frequency band. Its nonlinear layers are constructed using programmable metasurfaces integrated with RF components, implementing a ReLU-like activation function. The nonlinear architecture achieves a nanosecond-scale delay (17.7 ns), representing orders of magnitude improvement in speed over photoelectric conversion-based nonlinearities. Moreover, the nonlinearity is characterized by exceedingly low thresholds and reconfigurable nonlinear activation functions. The system demonstrates remarkable classification capability in image classification and real-time human posture recognition tasks. Characterized by low latency, high speed, low power consumption, and flexible nonlinear activation, this architecture holds great promise for applications in security screening, medical rehabilitation, human-computer interaction, and numerous other fields.
- 24 November 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 10332
Open AccessMultilayer nonlinear diffraction neural networks with programmable and fast ReLU activation function
Optical diffractive neural networks are emerging for improving speed and energy efficiency in machine learning. However, the challenges of nonlinear activation functions (e.g., latency issues, high power consumption, and cascading complexity) impede their performance and practical deployment. Here, we propose a programmable multilayer full-space nonlinear neural network operating in the microwave frequency band. Its nonlinear layers are constructed using programmable metasurfaces integrated with RF components, implementing a ReLU-like activation function. The nonlinear architecture achieves a nanosecond-scale delay (17.7 ns), representing orders of magnitude improvement in speed over photoelectric conversion-based nonlinearities. Moreover, the nonlinearity is characterized by exceedingly low thresholds and reconfigurable nonlinear activation functions. The system demonstrates remarkable classification capability in image classification and real-time human posture recognition tasks. Characterized by low latency, high speed, low power consumption, and flexible nonlinear activation, this architecture holds great promise for applications in security screening, medical rehabilitation, human-computer interaction, and numerous other fields.
- Open Access
- 1 July 2025
- Published by AIP Publishing in APL Photonics
- Vol. 10 (7)
Optimized dimension-reduction algorithm on input layer for diffractive neural networks
Diffractive neural network (DNN) has provided a novel solution to address complex artificial intelligence tasks due to its ultrahigh processing speed and parallel computing capability. However, traditional DNNs lack fine design of input layers, which limits their performance in miniaturized applications. Here, we propose an optimized dimension-reduction algorithm based on an improved Fisher score to directly filter the diffractive information injected into the network, enabling DNNs to obtain inputs with stronger resolving capabilities. Numerical results demonstrate that our approach achieves accuracy improvements of 5% and 7.3% on the MNIST and Fashion-MNIST classification datasets, respectively, compared to the uniform grid sampling method that is commonly used in the existing dimension-reduction DNNs. In addition, when addressing the issue of complex input image distortion, our method holds the accuracy advantage of 4.8% in the pincushion distorted MNIST dataset and expands it to 6.7% in the tangential distortion scenario. Our optimized dimension-reduction algorithm has effectively utilized the information in diffractive fields, providing an efficient solution to the miniaturization of DNNs and promoting their application prospects in wearable devices, industrial inspection, and autonomous driving systems.
- 1 July 2025
- Published by AIP Publishing in APL Photonics
- Vol. 10 (7)
Open AccessOptimized dimension-reduction algorithm on input layer for diffractive neural networks
Diffractive neural network (DNN) has provided a novel solution to address complex artificial intelligence tasks due to its ultrahigh processing speed and parallel computing capability. However, traditional DNNs lack fine design of input layers, which limits their performance in miniaturized applications. Here, we propose an optimized dimension-reduction algorithm based on an improved Fisher score to directly filter the diffractive information injected into the network, enabling DNNs to obtain inputs with stronger resolving capabilities. Numerical results demonstrate that our approach achieves accuracy improvements of 5% and 7.3% on the MNIST and Fashion-MNIST classification datasets, respectively, compared to the uniform grid sampling method that is commonly used in the existing dimension-reduction DNNs. In addition, when addressing the issue of complex input image distortion, our method holds the accuracy advantage of 4.8% in the pincushion distorted MNIST dataset and expands it to 6.7% in the tangential distortion scenario. Our optimized dimension-reduction algorithm has effectively utilized the information in diffractive fields, providing an efficient solution to the miniaturization of DNNs and promoting their application prospects in wearable devices, industrial inspection, and autonomous driving systems.
- Open Access
- 6 August 2024
- Published by Springer Nature in Nature Communications
- Vol. 15 (1) , 1-11
Terahertz spoof plasmonic neural network for diffractive information recognition and processing
All-optical diffractive neural networks, as analog artificial intelligence accelerators, leverage parallelism and analog computation for complex data processing. However, their low space transmission efficiency or large spatial dimensions hinder miniaturization and broader application. Here, we propose a terahertz spoof plasmonic neural network on a planar diffractive platform for direct multi-target recognition. Our approach employs a spoof surface plasmon polariton coupler array to construct a diffractive network layer, resulting in a compact, efficient, and easily integrable architecture. We designed three schemes: basis vector classification, multi-user recognition, and MNIST handwritten digit classification. Experimental results reveal that the terahertz spoof plasmonic neural network successfully classifies basis vectors, recognizes multi-user orientation information, and directly processes handwritten digits using a designed input framework comprising a metal grating array, transmitters, and receivers. This work broadens the application of terahertz plasmonic metamaterials, paving the way for terahertz on-chip integration, intelligent communication, and advanced computing systems.
- 6 August 2024
- Published by Springer Nature in Nature Communications
- Vol. 15 (1) , 1-11
Open AccessTerahertz spoof plasmonic neural network for diffractive information recognition and processing
All-optical diffractive neural networks, as analog artificial intelligence accelerators, leverage parallelism and analog computation for complex data processing. However, their low space transmission efficiency or large spatial dimensions hinder miniaturization and broader application. Here, we propose a terahertz spoof plasmonic neural network on a planar diffractive platform for direct multi-target recognition. Our approach employs a spoof surface plasmon polariton coupler array to construct a diffractive network layer, resulting in a compact, efficient, and easily integrable architecture. We designed three schemes: basis vector classification, multi-user recognition, and MNIST handwritten digit classification. Experimental results reveal that the terahertz spoof plasmonic neural network successfully classifies basis vectors, recognizes multi-user orientation information, and directly processes handwritten digits using a designed input framework comprising a metal grating array, transmitters, and receivers. This work broadens the application of terahertz plasmonic metamaterials, paving the way for terahertz on-chip integration, intelligent communication, and advanced computing systems.
- 6 June 2023
- Published by Wiley in Laser & Photonics Reviews
- Vol. 17 (11)
Reconfigurable Spoof Localized Surface Plasmonic for Frequency Detections
Under the impetus of wireless communications, metamaterials have attracted considerable attention due to their superior merits of free manipulation of electromagnetic waves and easy integration. However, most efforts to date are mainly focused on the modulation of wave information or recognition of wave intensity, while the frequency detection is rarely explored. Here, a spoof localized surface plasmon (SLSP) structure is proposed to realize reconfigurable frequency detections. By switching the bias voltage applied to the varactor, the resonance frequency of SLSP can be controlled dynamically from 1.95 to 2.35 GHz. The SLSP‐based waveguide has a relatively high quality factor due to the strongly localized field confinement. Reconfigurable frequency detection is demonstrated experimentally by monitoring the output voltage of the introduced detector. The measured and simulated results are in good agreement. This work will hold promising applications in wireless communication and sensing systems.
- 6 June 2023
- Published by Wiley in Laser & Photonics Reviews
- Vol. 17 (11)
Reconfigurable Spoof Localized Surface Plasmonic for Frequency Detections
Under the impetus of wireless communications, metamaterials have attracted considerable attention due to their superior merits of free manipulation of electromagnetic waves and easy integration. However, most efforts to date are mainly focused on the modulation of wave information or recognition of wave intensity, while the frequency detection is rarely explored. Here, a spoof localized surface plasmon (SLSP) structure is proposed to realize reconfigurable frequency detections. By switching the bias voltage applied to the varactor, the resonance frequency of SLSP can be controlled dynamically from 1.95 to 2.35 GHz. The SLSP‐based waveguide has a relatively high quality factor due to the strongly localized field confinement. Reconfigurable frequency detection is demonstrated experimentally by monitoring the output voltage of the introduced detector. The measured and simulated results are in good agreement. This work will hold promising applications in wireless communication and sensing systems.