- Ferroelectric and Negative Capacitance Devices
- Advanced Memory and Neural Computing
- Machine Learning and ELM
- Phonetics and Phonology Research
- linguistics and terminology studies
- Linguistic Variation and Morphology
- AI in cancer detection
- Semiconductor materials and devices
- Smart Agriculture and AI
- Multilingual Education and Policy
- Linguistics, Language Diversity, and Identity
- China's Ethnic Minorities and Relations
- Electrical Fault Detection and Protection
- Sentiment Analysis and Opinion Mining
- Advanced Data Storage Technologies
- Fluid Dynamics and Turbulent Flows
- Remote Sensing in Agriculture
- COVID-19 and healthcare impacts
- Combustion and Detonation Processes
- Magnetic Field Sensors Techniques
- Angiogenesis and VEGF in Cancer
- Granular flow and fluidized beds
- Neural Networks and Reservoir Computing
- Digital Imaging for Blood Diseases
- Long-Term Effects of COVID-19
South China University of Technology
2023
National University of Defense Technology
2018-2022
Huazhong University of Science and Technology
2012-2020
Union Hospital
2020
The University of Texas Southwestern Medical Center
2017-2018
Highland Community College - Kansas
2017
Wuhan National Laboratory for Optoelectronics
2012-2013
Tsinghua University
2007
Nous dressons la liste minimale des traits distinctifs nécessaires pour rendre compte oppositions vocaliques en qiang yadu, à partir d'une analyse mécanismes régissant l'harmonie vocalique. discutons arguments plaidant ou contre binante de ces traits, ainsi que certains contrastes marginaux et voie disparition. Cet inventaire minimal prédit l'existence voyelles abstraites, nous montrons le rôle joué par celles-ci dans système Les binaires sont [antérieur], [ATR], [bas], tandis les [arrondi]...
Pathological angiogenesis has been identified in many malignancies as a potential prognostic factor and target for therapy. In most cases, angiogenic analysis is based on the measurement of microvessel density (MVD) detected by immunostaining CD31 or CD34. However, retrievable public data generally composed Hematoxylin Eosin (H&E)-stained pathology images, which difficult to get corresponding immunohistochemistry images. The role microvessels H&E stained images not widely studied due their...
Resistive switching random access memory (RRAM) shows its potential to be a promising candidate as the basic in-memory computing unit for deep neural networks (DNN) accelerator design due non-volatile, low power, and small footprint properties. The RRAM based crossbar array (RRAM CBA) is usually employed accelerate DNN because of intrinsic characteristic executing multiplication-and-accumulation (MAC) operation according Kirchhoffs' law. However, some major non-ideal effects including...
Biomass is an important component of the plant phenomics, and existing methods for biomass estimation individual plants are either destructive or lack accuracy. In this study, a hyperspectral imaging system was developed accurate prediction above-ground rice in visible near-infrared spectral region. First, structure influence various parameters on camera acquisition speed were established. Then used to image 152 plants, which selected from mini-core collection, two stages, tillering...
Background Coronavirus disease 2019 (COVID-19) has now brought major challenges to public health and the economy globally since December 2019, which requires effective treatment prevention strategies adapt impact of pandemic. We, therefore, explored prognostic factors for patients with COVID-19 contribution immunomodulatory therapy on outcome.
The unwanted electro-forming process is unavoidable for the practical application of most resistive random access memory (ReRAM) devices, which always being one obstacles massive commercialization this novel electronic device. In letter, a forming-free Pt/Ti/(TiO2-x)/Ta2O5/Pt based ReRAM device demonstrated with an additional feature low operation voltage. fitting result measured I-V curves reveals that switching fabricated conducted by electrons trapping/de-trapping in deep-level electron...
Resistive random-access memory (ReRAM)-based crossbar array (RCA) is a promising platform to accelerate vector-matrix multiplication in deep neural networks (DNNs). There are, however, some practical issues, especially device variation, that hinder the versatile development of ReRAM computing systems. The variations include device-to-device variation (DDV) and cycle-to-cycle (CCV) deviate devise resistance RCA from their target state. Such deviation seriously degrades inference accuracy...
In this paper, we propose a new 3-bit ADC based on magnetoelectric spin–orbit (MESO) device. By adjusting the thickness of part MESOs, possesses multiple switching thresholds for converting input voltages into magnetization state. means iterative HSPICE analysis and Monte Carlo analysis, quantitatively calculated power consumption evaluated reliability The results show that proposed could sample at an 8.3 GHz rate with 0.014 pJ conversion energy. comparison NAND-SPIN magnetic (MADC),...
Examining vowel harmony processes at work in Yadu Qiang reveals the smallest set of features needed to account for oppositions, hence structure. Arguments are evaluated binarity/unarity features, and roles marginal disappearing contrasts. A minimal that accounts predicts existence abstract vowels, which shown be active system. The required binary Front, TR, Low, while Round Rhotic unary features. study draws paralleis with Hyman's (2003) analysis Kalong, all unary.
With high memory density, non-volatility and DRAM-scale latency, non-volatile (NVM) brings evolution to storage systems durable data structures. And Intel Optane DC persistent module (AEP), the first commercial product of NVM, shows some features that are different from previous assumptions: higher read lower bandwidth block access granularity compared with DRAM. It is reasonable build up hybrid give full play complementary advantages DRAM NVM. In this paper, we present a read-efficient...
Traditional CMOS-based von-Neumann computer architecture faces the issue of memory wall that limitation bus-bandwidth and speed mismatch between processor restrict efficiency data processing along with an irreducible energy consumption conducted by movement, especially in some data-intensive applications. Recently, novel in-memory computing (IMC) paradigms developed utilizing characteristics different non-volatile memories provide promising ways to overcome bottleneck wall. Here, we propose...
This paper discusses the copula and existential verb constructions in Qiang, a Tibeto-Burman language of northern Sichuan, China. There is only one which can be used equational, identificational, attributive, naming, cleft constructions, as well type possessive construction. are five verbs use depends on semantics referent being predicated existing its location. The have number characteristics adjective-like stative verbs, modified by adverbs degree, but they cannot directly modify nouns....
Abstract Memristor stateful logic based on a two-dimensional (2D) memristor crossbar array (MCBA) has been proved to be an effective approach break the von Neumann bottleneck. But three-dimensional (3D) which inherent high storage density is less investigated in computation field. In this work, we implemented whole linearly separable binary Boolean functions completely by using circuits configured multi-layer array. This circuit primitive can execute multi-gates parallelly inside layer or...
Emerging resistive random-access memory (ReRAM) has demonstrated great potential in the achievement of in-memory computing paradigm to overcome well-known “memory wall” current von Neumann architecture. The ReRAM crossbar array (RCA) is a promising circuit structure accelerate vital multiplication-and-accumulation (MAC) operations deep neural networks (DNN). However, due nonlinear distribution conductance levels ReRAM, large deviation exists mapping process when trained weights that are...
Resistive Random Access Memory (RRAM) is a promising technology for efficient neural computing systems with low power, non-volatile, and good compatibility CMOS. The RRAM based crossbar usually employed to accelerate deep networks (DNN) because of its intrinsic characteristic executing multiplication-and-accumulation (MAC) operation according Kirchhoff's law. However, there are some realistic device issues especially the variation, stochastic behavior devices, resulting in significant...
In the past few years, there has been a significant increase in high-voltage cable joint breakdown and explosion incidents, resulting combustion situations deserve attention. This paper conducts an analysis of 220kV integrated prefabricated incident, utilizes finite element simulation software to simulate fluid field at site faulty joint, analyze impact location on situation joint. The results indicate that experiences bursting insulation flange exhibits intense due strong airflow...