- Force Microscopy Techniques and Applications
- Biochemical and Structural Characterization
- Evolutionary Game Theory and Cooperation
- Bone Tissue Engineering Materials
- Advanced Materials and Mechanics
- Plant and animal studies
- Nanopore and Nanochannel Transport Studies
- Speech Recognition and Synthesis
- Acoustic Wave Resonator Technologies
- Adhesion, Friction, and Surface Interactions
- Titanium Alloys Microstructure and Properties
- Music and Audio Processing
- Chemical Synthesis and Analysis
- Advanced Neural Network Applications
- Machine Learning in Materials Science
- Human Pose and Action Recognition
- Peptidase Inhibition and Analysis
- Microfluidic and Bio-sensing Technologies
- Structural Analysis and Optimization
- Advanced Vision and Imaging
- DNA and Nucleic Acid Chemistry
- Electrostatics and Colloid Interactions
- Antimicrobial Peptides and Activities
- Music History and Culture
- Soil Mechanics and Vehicle Dynamics
Collaborative Innovation Center of Advanced Microstructures
2025
Nanjing University
2025
City University of Hong Kong
2020-2024
Shandong University of Technology
2024
City University of Hong Kong, Shenzhen Research Institute
2020-2022
Nanjing University of Science and Technology
2018
Dental implants with long-term success of osseointegration have always been the goal, however, difficulties exist. The accumulation fretting damage at implant-bone interface often gets overlooked. Commonly used titanium is approximately 7-fold harder and stiffer than cortical bone. Stress shielding caused by mismatching elastic modulus aggravates interface, which accompanied risk formation proinflammatory metal debris implant loosening. Thus, authors explore functionalized bone-inspired...
Amino acid crystals have emerged as promising piezoelectric materials for biodegradable and biocompatible sensors; however, their relatively low coefficients constrain practical applications. Here, we introduce a fluoro‐substitution strategy to overcome this limitation enhance the performance of amino crystals. Specifically, substituted hydrogen atoms on aromatic rings L‐tryptophan, L‐phenylalanine, N‐Cbz‐L‐phenylalanine with fluorine, resulting in significantly elevated coefficients....
Tensegrity structures are isolated rigid compression components held in place by a continuous network of tensile components, and central to natural systems such as the extracellular matrix cell cytoskeleton. These enable nonreciprocal mechanical properties essential for dynamic biological functions. Here, we introduce synthetic approach engineer hydrogels with tensegrity architectures, drawing inspiration from mechanochemical principles underlying systems. By employing in-situ enzyme-induced...
Adversarial examples are hot topics in the field of security deep learning. The feature, generation methods, attack and defense methods adversarial focuses current research on examples. This article explains key technologies theories from concept examples, occurrences attacking lists possible reasons for also analyzes several typical detail: Limited-memory BFGS (L-BFGS), Fast Gradient Sign Method (FGSM), Basic Iterative (BIM), Least-likely Class (LLC), etc. Furthermore, perspective main...
Abstract Nano-hydroxyapatite (nano-HA) has attracted substantial attention in the field of regenerative medicine. Endothelial cell (EC)-mesenchymal stem (MSC) interactions are necessary for bone reconstruction, but manner which nano-HA interacts this process remains unknown. Herein, we investigated cytotoxicity and osteoinductive effects HA nanoparticles (HANPs) on MSCs using an indirect co-culture model mediated by ECs highlighted underlying mechanisms. It was found that at a subcytotoxic...
Knotting is common in long polymers and significantly affects the polymer behavior. It generally believed that knotting should increase bending energy. However, many recent studies found can surprisingly decrease energy, while reason remains unclear. In this work, we quantitatively explained surprising phenomenon using tube model. model, segments a knot core are confined virtual tube. The of knots two opposite ways. First, curved to assume knotted shape, which increases bending. Second,...
Knots can spontaneously form in polymers. How knotting affects polymer behavior depends on probability, pknot. An intriguing result about pknot recent studies is that exhibits a non-monotonic dependence the bending stiffness and maximized at Lp ≈ 8a, where persistence length hardcore diameter of monomer. In this work, we propose new explanation for based fact knots are typically localized. We find results from competition special entropic effect arising variation sizes localized an...
Knots can spontaneously form in DNA, proteins, and other polymers affect their properties. These knots often experience spatial confinement biological systems experiments. While dramatically affects the knot behavior, physical mechanisms underlying effects are not fully understood. In this work, we provide a simple picture of polymer slit using tube model. model, segments core assumed to be confined virtual due topological restriction. We first perform Monte Carlo simulation flexible knotted...
Double-helix structures, such as DNA, are formed in nature to realize many unique functions. Inspired by this, researchers pursuing strategies design structures from polymers. A key question is whether the double helix can be self-folding of a single polymer chain without specific interactions. Here, using Langevin dynamics simulation and theoretical analysis, we find that stable double-helix phase achieved semiflexible polymers result cooperation between local structure nonlocal attraction....
Extraordinarily stable protein and peptide structures are critically demanded in many applications. Typical approaches to enhance stability strengthening certain interactions. Here, we develop a very different approach: stabilizing through side-chain-locked knots. More specifically, core consists of knot, which is prevented from unknotting unfolding by large side chains amino acids at knot boundaries. These impose free energy barriers for unknotting. The quantified using all-atom...
The population dynamics of two interacting species modeled by the Lotka-Volterra (LV) model with general parameters that can promote or suppress other is studied. It found properties species' isoclines determine interaction species, leading to six regimes in phase diagram interspecies interaction; i.e., there are different interspecific relationships described LV model. Four allow for nontrivial coexistence, among which it three them stable, namely, weak competition, mutualism, and...
Action recognition in videos poses a challenge due to its high computational cost, especially for Joint Space-Time video transformers (Joint VT). Despite their effectiveness, the excessive number of tokens such architectures significantly limits efficiency. In this paper, we propose HaltingVT, an efficient transformer adaptively removing redundant patch tokens, which is primarily composed VT and Glimpser module. Specifically, HaltingVT applies data-adaptive token reduction at each layer,...
The performance of streaming automatic speech recognition (ASR) is often inferior to that non-streaming due the absence complete contextual information. However, we cannot optimize model by merely accessing more future frames, as this would lead considerable latency. In paper, propose Future-aware Transformer (FaT) models long-distance dependencies transferring information from later chunks former through look-ahead windows. Specifically, chunk-based context used encode audio sequence...
In this study, a closed multi-channel air-blowing plug seedling pick-up device and combined tray were designed to address the issues of complex structure, high damage rates low efficiency in fully automated vegetable transplanter systems. The operates by sealing seedlings cup, where compressed air is channeled into sealed cavity through multiple passages during process. upper surface subjected uniform force, overcoming friction adhesion between tray. This process presses guide tube,...
In this work, we report an intriguing phenomenon: crowding-induced polymer trapping in a channel. Using Langevin dynamics simulations and analytical calculations, find that for confined channel, crowding particles can push into the channel corner through inducing effective polymer-corner attraction due to depletion effect. This phenomenon is referred as trapping. The occurrence of requires minimum volume fraction crowders, ϕ^{*}, which scales ϕ^{*}∼(a_{c}/L_{p})^{1/3} a_{c}≫a_{m}...
To achieve lightweight object detectors for deployment on the edge devices, an effective model compression pipeline is proposed in this paper. The consists of automatic channel pruning backbone, fixed deletion branch layers and knowledge distillation guidance learning. As results, Resnet50-v1d auto-pruned fine-tuned ImageNet to attain a compact base as backbone detector. Then, are implemented with pipeline. For instance, SSD-300 size=16.3MB, FLOPS=2.31G, mAP=71.2 created, revealing better...
A large number of pornographic audios publicly available on the Internet seriously threaten mental and physical health children, but these are rarely detected filtered. In this paper, we firstly propose a convolutional neural networks (CNN) based model for acoustic pornography recognition. Then, research collection refinements verify their effectiveness through ablation studies. Finally, stack all together to whether they can further improve accuracy model. Experimental results our...