Nanjing University of Science and Technology

About: Nanjing University of Science and Technology is a education organization based out in Nanjing, China. It is known for research contribution in the topics: Control theory & Catalysis. The organization has 31581 authors who have published 36390 publications receiving 525474 citations. The organization is also known as: Nánjīng Lǐgōng Dàxué & Nánlǐgōng.

Selective Kernel Networks

Abstract: In standard Convolutional Neural Networks (CNNs), the receptive fields of artificial neurons in each layer are designed to share the same size. It is well-known in the neuroscience community that the receptive field size of visual cortical neurons are modulated by the stimulus, which has been rarely considered in constructing CNNs. We propose a dynamic selection mechanism in CNNs that allows each neuron to adaptively adjust its receptive field size based on multiple scales of input information. A building block called Selective Kernel (SK) unit is designed, in which multiple branches with different kernel sizes are fused using softmax attention that is guided by the information in these branches. Different attentions on these branches yield different sizes of the effective receptive fields of neurons in the fusion layer. Multiple SK units are stacked to a deep network termed Selective Kernel Networks (SKNets). On the ImageNet and CIFAR benchmarks, we empirically show that SKNet outperforms the existing state-of-the-art architectures with lower model complexity. Detailed analyses show that the neurons in SKNet can capture target objects with different scales, which verifies the capability of neurons for adaptively adjusting their receptive field sizes according to the input. The code and models are available at this https URL.

Improving All‐Inorganic Perovskite Photodetectors by Preferred Orientation and Plasmonic Effect

Abstract: All-inorganic perovskites have high carrier mobility, long carrier diffusion length, excellent visible light absorption, and well overlapping with localized surface plasmon resonance (LSPR) of noble metal nanocrystals (NCs). The high-performance photodetectors can be constructed by means of the intrinsic outstanding photoelectric properties, especially plasma coupling. Here, for the first time, inorganic perovskite photodetectors are demonstrated with synergetic effect of preferred-orientation film and plasmonic with both high performance and solution process virtues, evidenced by 238% plasmonic enhancement factor and 106 on/off ratio. The CsPbBr3 and Au NC inks are assembled into high-quality films by centrifugal-casting and spin-coating, respectively, which lead to the low cost and solution-processed photodetectors. The remarkable near-field enhancement effect induced by the coupling between Au LSPR and CsPbBr3 photogenerated carriers is revealed by finite-difference time-domain simulations. The photodetector exhibits a light on/off ratio of more than 106 under 532 nm laser illumination of 4.65 mW cm−2. The photocurrent increases from 0.67 to 2.77 μA with centrifugal-casting. Moreover, the photocurrent rises from 245.6 to 831.1 μA with Au NCs plasma enhancement, leading to an enhancement factor of 238%, which is the most optimal report among the LSPR-enhanced photodetectors, to the best of our knowledge. The results of this study suggest that all-inorganic perovskites are promising semiconductors for high-performance solution-processed photodetectors, which can be further enhanced by Au plasmonic effect, and hence have huge potentials in optical communication, safety monitoring, and biological sensing.

Observer-Based Adaptive Neural Network Control for Nonlinear Stochastic Systems With Time Delay

Abstract: This paper considers the problem of observer-based adaptive neural network (NN) control for a class of single-input single-output strict-feedback nonlinear stochastic systems with unknown time delays. Dynamic surface control is used to avoid the so-called explosion of complexity in the backstepping design process. Radial basis function NNs are directly utilized to approximate the unknown and desired control input signals instead of the unknown nonlinear functions. The proposed adaptive NN output feedback controller can guarantee all the signals in the closed-loop system to be mean square semi-globally uniformly ultimately bounded. Simulation results are provided to demonstrate the effectiveness of the proposed methods.

Controllable synthesis of porous Fe3O4@ZnO sphere decorated graphene for extraordinary electromagnetic wave absorption

Abstract: For the first time, mesoporous Fe3O4@ZnO sphere decorated graphene (GN–pFe3O4@ZnO) composites with uniform size, considerable porosity, high magnetization and extraordinary electromagnetic (EM) wave absorption properties were synthesized by a simple and efficient three-step method. Structure and morphology details were characterized by X-ray diffraction, transmission electron microscopy, high-resolution electron microscopy and field-emission scanning electron microscopy. Electron microscopy images reveal that pFe3O4@ZnO spheres with obvious porous and core–shell structures are uniformly coated on both sides of the GN sheets without significant numbers of vacancies or apparent aggregation. EM wave absorption properties of epoxy containing 30 wt% GN–pFe3O4@ZnO were investigated at room temperature in the frequency region of 0.2–18 GHz. The absorption bandwidth with reflection loss (RL) values less than −10 dB is up to 11.4 GHz, and the minimal RL is almost −40 dB. The intrinsic physical and chemical properties of the materials, the synergy of Fe3O4 and ZnO, and particularly the unique multi-interfaces are fundamental to the enhancement of EM absorption properties. The as-prepared GN–pFe3O4@ZnO composites are shown to be lightweight, have strong absorption, and broad frequency bandwidth EM absorbers.