A5071230229- Magnetic properties of thin films
- Speech Recognition and Synthesis
- Natural Language Processing Techniques
- Physics of Superconductivity and Magnetism
- Magnetic and transport properties of perovskites and related materials
- Rare-earth and actinide compounds
- Topic Modeling
- Advanced Memory and Neural Computing
- Ferroelectric and Negative Capacitance Devices
- Iron-based superconductors research
- Theoretical and Computational Physics
- Advanced Condensed Matter Physics
- Characterization and Applications of Magnetic Nanoparticles
- Metamaterials and Metasurfaces Applications
- Electronic and Structural Properties of Oxides
- Network Traffic and Congestion Control
- Anodic Oxide Films and Nanostructures
- Atomic and Subatomic Physics Research
- Quantum and electron transport phenomena
- Music and Audio Processing
- Identification and Quantification in Food
- Marine and fisheries research
- Magnetic Properties of Alloys
- Chalcogenide Semiconductor Thin Films
- Supercapacitor Materials and Fabrication
The People's Hospital of Guangxi Zhuang Autonomous Region
2024-2025
National University of Singapore
2009-2025
Tianjin University
2025
Institute of Materials Research and Engineering
1999-2024
Agency for Science, Technology and Research
2015-2024
Southern Medical University
2023
East China University of Technology
2023
Zhangzhou Normal University
2022
Data Storage Institute
2015-2022
University of Cambridge
1991-2020
In recent years recurrent neural network language models (RNNLMs) have been successfully applied to a range of tasks including speech recognition. However, an important issue that limits the quantity data used, and their possible application areas, is computational cost in training. A signi??cant part this associated with softmax function at output layer, as requires normalization term be explicitly calculated. This impacts both training testing speed, especially when large vocabulary used....
In recent years, recurrent neural network language models (RNNLMs) have become increasingly popular for a range of applications including speech recognition. However, the training RNNLMs is computationally expensive, which limits quantity data, and size network, that can be used. order to fully exploit power RNNLMs, efficient implementations are required. This paper introduces an open-source toolkit, CUED-RNNLM supports GPU-based RNNLMs. RNNLM with large number word level output targets...
Magnetic transition metal dichalcogenide (TMD) films have recently emerged as promising candidates to host novel magnetic phases relevant next-generation spintronic devices. However, systematic control of the magnetization orientation, or anisotropy, and its thermal stability, characterized by Curie temperature (Tc) remains be achieved in such films. Here we present self-intercalated epitaxial Cr1+{\delta}Te2 a platform for achieving systematic/smooth tailoring TMD Using molecular beam...
Controllable writing and deleting of nanoscale magnetic skyrmions are key requirements for their use as information carriers next-generation memory computing technologies. While several schemes have been proposed, they require complex fabrication techniques or precisely tailored electrical inputs, which limits long-term scalability. Here we demonstrate an alternative approach using conventional pulses within a simple, two-terminal wire geometry. X-ray microscopy experiments micromagnetic...
Antiferromagnetic (AF) skyrmions are topological spin structures with fully compensated, net-zero magnetization. Compared to their ferromagnetic (FM) skyrmion counterparts, reduced stray field and enhanced electrical response can enable linear, high-throughput current-driven motion. However, bubble-like character in conventional bilayer AFs limits stability fluctuations, leading deformation annihilation. Here we present the engineering of a composite AF chiral multilayer, wherein interplay...
Abstract Renal fibrosis is a common mechanism leading to kidney failure in chronic diseases (CKDs), including obstructive nephropathy (ON). Dysregulated inflammation central the development of renal fibrosis, but how local immune cells within tissue microenvironment integrate and coordinate drive this condition remains largely unknown. Herein, we documented that neutrophils were abundantly recruited expelled neutrophil extracellular traps (NETs) human mouse fibrotic kidneys. Importantly,...
A facile, mild, and scalable late-stage N-alkenylative modification strategy is introduced on 1H-indoles, 9H-carbazoles, their structural derivatives analogues, including alkaloids, bioactive agents, tryptophan motifs, via chemo- regioselective phosphine-mediated propiolate hydroamination. Saliently, through this protocol, bisconjugation macrocyclization (bis)indolic scaffolds can also be accomplished, with the installation of new α,β-unsaturated ester handles for potential further versatile...
Recurrent neural network language models (RNNLMs) are becoming increasingly popular for speech recognition. Previously, we have shown that RNNLMs with a full (non-classed) output layer (F-RNNLMs) can be trained efficiently using GPU giving large reduction in training time over conventional class-based (C-RNNLMs) on standard CPU. However, since test-time RNNLM evaluation is often performed entirely CPU, F-RNNLMs inefficient the entire needs to calculated normalisation. In this paper, it...
Abstract One early triumph of quantum physics is the explanation why some materials are metallic whereas others insulating. While a treatment based on single electron states correct for most this approach can fail spectacularly, when electrostatic repulsion between electrons causes strong correlations. Not only these favor new and subtle forms matter, such as magnetism or superconductivity, they even cause in half-filled energy band to lock into position, producing correlated, Mott...
The realization of chiral spin textures, comprising myriad distinct, nanoscale arrangements spins with topological properties, has established pathways for engineering robust, energy-efficient, and scalable elements non-volatile nanoelectronics. Particularly, current-induced manipulation textures in nanowire racetracks tunnel junction based devices are actively investigated applications memory, logic, unconventional computing. In this Article, we paint a background on the progress as well...
Objectives/Hypothesis Paradoxical vocal fold motion and exercise‐induced paradoxical (EIPVFM) are two related conditions that do not have definitive diagnostic criteria. Much of the EIPVFM literature describes patients with characteristic physiologic findings severe upper airway obstruction or obvious airflow limitation in clinical context exertional dyspnea audible stridor. The objective this study was to highlight a group who demonstrate important (exertional stridor) without...
Magnetic skyrmions are topologically wound nanoscale textures of spins whose ambient stability and electrical manipulation in multilayer films have led to an explosion research activities. While past efforts focused predominantly on isolated skyrmions, recently ensembles chiral spin textures, consisting magnetic stripes, shown possess rich interactions with potential for device applications. However, several fundamental aspects texture phenomenology remain be elucidated, including their...
Abstract Magnetic skyrmions have emerged as promising elements for encoding information toward biomimetic computing applications due to their pseudoparticle nature and efficient coupling spin currents. A key hindrance skyrmionic devices is instability against elongation at zero magnetic field (ZF). Prevailing materials approaches focused on tailoring skyrmion energetics found ZF configurations be highly sensitive, which imposes significant growth constraints limits device scalability. This...
The intake of dietary antioxidants and glycolipid metabolism are closely related to chronic kidney disease (CKD), particularly among individuals with abdominal obesity. Nevertheless, the cumulative effect multiple comorbid risk factors on progression complications CKD remains inadequately characterized. This study analyzed data from National Health Nutrition Examination Survey (NHANES) dat abase (2005–2018), examine potential CKD, including metabolism, antioxidant intake, pertinent medical...
State-of-the-art English automatic speech recognition systems typically use phonetic rather than graphemic lexicons. Graphemic are known to perform less well for as the mapping from written form spoken is complicated. However, in recent years representational power of deep-learning based acoustic models has improved, raising interest English, due simplicity generating lexicon. In this paper, and compared an Multi-Genre Broadcast transcription task. A range on lattice-free MMI training...
Magnetic skyrmions form in chiral multilayers from the shrinking or fission of elongated stripe textures. Here we report an experimental and theoretical study temperature dependence this stripe-to-skyrmion transition $\mathrm{Co}$/$\mathrm{Pt}$-based multilayers. Field-reversal magnetometry Lorentz microscopy experiments over 100--350 K establish increased efficacy at higher temperatures---driven primarily by thermal evolution key magnetic interactions---which enhances resulting skyrmion...
The symmetric (Heisenberg) exchange interaction is fundamental to magnetism and assumes critical importance in designing magnetic materials for novel emergent phenomena device applications. However, quantifying extremely challenging ultrathin films, as techniques approximations reliably used bulk are largely inapplicable the two-dimensional limit. Here we present contrast measurement of stiffness $A$ by several methods on a series five Co/Pt-based ($1--2$ nm) films. We compare results from...