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.

Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection

Abstract: One-stage detector basically formulates object detection as dense classification and localization. The classification is usually optimized by Focal Loss and the box location is commonly learned under Dirac delta distribution. A recent trend for one-stage detectors is to introduce an individual prediction branch to estimate the quality of localization, where the predicted quality facilitates the classification to improve detection performance. This paper delves into the representations of the above three fundamental elements: quality estimation, classification and localization. Two problems are discovered in existing practices, including (1) the inconsistent usage of the quality estimation and classification between training and inference and (2) the inflexible Dirac delta distribution for localization when there is ambiguity and uncertainty in complex scenes. To address the problems, we design new representations for these elements. Specifically, we merge the quality estimation into the class prediction vector to form a joint representation of localization quality and classification, and use a vector to represent arbitrary distribution of box locations. The improved representations eliminate the inconsistency risk and accurately depict the flexible distribution in real data, but contain continuous labels, which is beyond the scope of Focal Loss. We then propose Generalized Focal Loss (GFL) that generalizes Focal Loss from its discrete form to the continuous version for successful optimization. On COCO test-dev, GFL achieves 45.0\% AP using ResNet-101 backbone, surpassing state-of-the-art SAPD (43.5\%) and ATSS (43.6\%) with higher or comparable inference speed, under the same backbone and training settings. Notably, our best model can achieve a single-model single-scale AP of 48.2\%, at 10 FPS on a single 2080Ti GPU. Code and models are available at this https URL.

Amino-Mediated Anchoring Perovskite Quantum Dots for Stable and Low-Threshold Random Lasing

Abstract: Halide perovskite quantum dots (Pe-QDs) have been considered as outstanding candidates for photodetector, light-emitting diode, and lasing applications, but these perspectives are being impeded by the severe stability, including both chemical and optical degradations. This study reports on amino-mediated anchoring Pe-QDs onto the surfaces of monodisperse silica to effectively depress the optical degradation of their photoluminescence (PL) and random lasing stabilities, hence achieving highly stable and low-threshold lasing. An amination-mediated nucleation and growth process is designed for the general and one-pot synthesis of Pe-QDs on the surfaces of silica spheres. The facile synthetic process, which can be finished within several minutes, insures scalable production. Surprisingly, almost no PL degradation is observed after 40 d storage under ambient conditions, even 80% PL intensity can be maintained after persistently illuminated by UV lamps for 108 h. Subsequently, extremely stable random lasing is achieved after storage for 2 months or over continuously optical pumping for 8 h. Such high PL and lasing stabilities originate from the isolation effects due to the effective anchoring, which separate the Pe-QDs from each other and inhibit the photoinduced regrowth and deterioration. This work will also open the window of perovskite-based multifunctional systems.

Robust H/sub /spl infin// control for uncertain discrete-time-delay fuzzy systems via output feedback controllers

Abstract: This work investigates the problem of robust output feedback H/sub /spl infin// control for a class of uncertain discrete-time fuzzy systems with time delays. The state-space Takagi-Sugeno fuzzy model with time delays and norm-bounded parameter uncertainties is adopted. The purpose is the design of a full-order fuzzy dynamic output feedback controller which ensures the robust asymptotic stability of the closed-loop system and guarantees an H/sub /spl infin// norm bound constraint on disturbance attenuation for all admissible uncertainties. In terms of linear matrix inequalities (LMIs), a sufficient condition for the solvability of this problem is presented. Explicit expressions of a desired output feedback controller are proposed when the given LMIs are feasible. The effectiveness and the applicability of the proposed design approach are demonstrated by applying this to the problem of robust H/sub /spl infin// control for a class of uncertain nonlinear discrete delay systems.

Fundamentals of superior properties in bulk NanoSPD materials

Abstract: Bulk nanoSPD materials are materials with nanostructural features, such as nanograins, nanoclusters, or nanotwins, produced by severe plastic deformation (SPD) techniques. Such nanostructured materials are fully dense and contamination free and in many cases they have superior mechanical and functional properties. Here, we provide a critical overview of such materials, with a focus on the fundamentals for the observed extraordinary properties. We discuss the unique nanostructures that lead to the superior properties, the underlying deformation mechanisms, the critical issues that remain to be investigated, future research directions, and the application potential of such materials.