A5107748550- AI in cancer detection
- Stock Market Forecasting Methods
- Advanced Graph Neural Networks
- Complex Systems and Time Series Analysis
- Glycosylation and Glycoproteins Research
- Cell Image Analysis Techniques
- Lignin and Wood Chemistry
- Machine Learning and Data Classification
- Topic Modeling
- Financial Markets and Investment Strategies
- Energy Load and Power Forecasting
- Catalysis for Biomass Conversion
- Carbohydrate Chemistry and Synthesis
- Biofuel production and bioconversion
- Advanced Neural Network Applications
- Thermal Radiation and Cooling Technologies
- Advanced SAR Imaging Techniques
- Thermal Expansion and Ionic Conductivity
- Regional Economic and Spatial Analysis
- Aviation Industry Analysis and Trends
- Seaweed-derived Bioactive Compounds
- Optical properties and cooling technologies in crystalline materials
- Radar Systems and Signal Processing
- Bauxite Residue and Utilization
- Advanced Memory and Neural Computing
Peking University
2021-2025
Institute of Basic Medical Sciences of the Chinese Academy of Medical Sciences
2025
University of South China
2025
Chinese Institute for Brain Research
2025
Guangzhou Automobile Group (China)
2024
South China University of Technology
2024
Henan University of Technology
2022-2024
Guangdong University of Technology
2022-2024
University of San Francisco
2020-2024
Donghua University
2023
Traditional object detection methods usually underperform when locating tiny or small drones against complex backgrounds, since the appearance features of targets and backgrounds are highly similar. To address this, inspired by magnocellular motion processing mechanisms, we proposed to utilize spatial–temporal characteristics flying based on spiking neural networks, thereby developing Magno-Spiking Neural Network (MG-SNN) for drone detection. The MG-SNN can learn identify potential regions...
Stock trend prediction is a challenging task due to the non-stationary dynamics and complex market dependencies. Existing methods usually regard each stock as isolated for prediction, or simply detect their correlations based on fixed predefined graph structure. Genuinely, associations stem from diverse aspects, underlying relation signals should be implicit in comprehensive graphs. On other hand, RNN network mainly used model historical data, while hard capture fine-granular volatility...
Stock investment selection is a hard issue in the Fintech field due to non-stationary dynamics and complex market interdependencies. Existing studies are mostly based on RNNs, which struggle capture interactive information among fine granular volatility patterns. Besides, they either treat stocks as isolated, or presuppose fixed graph structure heavily relying prior domain knowledge. In this paper, we propose novel Adaptive Long-Short Pattern Transformer (ALSP-TF) for stock ranking terms of...
Diol-modulated acidic hydrotropes gently extract native-like lignin while preserving cellulose, unlocking efficient biofuel and sunscreen production from diverse biomass.
A novel concept is introduced for fractionating biomass, specifically softwood, by employing a unique strategy that combines chemical protection and mitigation of lignin spatial confinement during the extraction process.
This study aimed to explore hyper-O-linked N-acetylglucosaminylation (O-GlcNAcylation) with an elevation of the expression O-linked-β-N-acetylglucosamine transferase (OGT) in human bladder cancer.Immunohistochemical staining for OGT and O-GlcNAcylation was performed 20 paired cancer adjacent normal tissues, as well tissue microarrays (N = 169). The level cell lines were detected using Western blot analysis. effects on proliferation 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide...
Out-of-distribution (OOD) graph generalization are critical for many real-world applications. Existing methods neglect to discard spurious or noisy features of inputs, which irrelevant the label. Besides, they mainly conduct instance-level class-invariant learning and fail utilize structural class relationships between instances. In this work, we endeavor address these issues in a unified framework, dubbed <bold xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Passive radiative cooling technology without electric consumption is an emerging sustainability that plays a key role in advancing sustainable development. However, most materials are vulnerable to outdoor contamination and thermal/UV exposure, which leads decreased performance. Herein, we report hierarchically structured polyimide/zinc oxide (PINF/ZnO) composite membrane integrates sunlight reflectance of 91.4% the main thermal effect solar spectrum (0.78-1.1 μm), mid-infrared emissivity...
Bitter peptides in the enzymatic hydrolysates were prepared and purified from wheat gluten using aqueous ethanol solutions macroporous resin, which has opened a new road for extraction separation of bitter peptides. This report contains release regularity factors affecting change intensity during hydrolysis, providing scientific basis research on debitterizing method. In this study, effects different degrees hydrolysis (DH) enzyme active sites peptide content taste thresholds discussed. The...
This study investigates the impact of science and technology talent policy on concentration talents in Guangdong, Hong Kong, Macao Greater Bay Area from 2010 to 2017. Utilizing data China Urban Statistical Yearbook various prefecture-level city statistical yearbooks, employs a combination single-difference double-difference methods analyze effects policy. The results indicate that significantly enhances region, with method providing more accurate estimates than method. research further...
Anoctamin/TMEM16 family members have recently been identified as novel calcium-activated chloride channels, and dysregulation of many participates in tumorigenesis progression. However, the exact role anoctamin5 (ANO5), one member this family, thyroid cancer is still not clarified. In study, we firstly found that expression levels ANO5 was significantly downregulated compared to adjacent normal tissue by mining public GEO database. Subsequently, further demonstrated 69.5% (57/82) clinical...
Stock recommendation plays a critical role in modern quantitative trading. The large volumes of social media information such as investment reviews that delegate emotion-driven factors, together with price technical indicators formulate “snapshot” the evolving stock market profile. However, previous studies usually model temporal trajectories and modalities separately while losing their interrelated influences. Moreover, they mainly extract review semantics via sequential or attentive...
Smart homes of the future are envisioned to have ability recognize many types home activities such as running a washing machine, flushing toilet, and using microwave. In this paper, we present new sensing technology, VibroSense, which is able 18 different throughout house by observing structural vibration patterns on wall or ceiling laser Doppler vibrometer. The received data processed sent deep neural network trained distinguish between activities. We conducted system evaluation, where...
Stock forecast is a crucial yet challenging task in modern quantitative trading. Given theoretical and investment merits, recently variety of deep learning methods have been proposed for automatically simulating stock movements from historical time series. However, these typically follow the i.i.d. assumption that actually contradicts complex trading environment. In reality, individual stocks often exhibit diverse volatility patterns, while macro market scenarios may also change over time,...
Abstract A tetrasaccharide repeating unit of the O ‐antigen from Escherichia coli O33 has been synthesized for first time using a chemoselective [1+1+2] one‐pot glycan assembly strategy. The phosphoglycerol moiety was installed by phosphoramidite method with benzylidene‐protected glycerol 2‐phosphoramidite as reagent. Overall, target monosaccharide building blocks in 12 linear steps, resulting satisfactory yield 28 %.
The valuation of large variable annuity portfolios is an important enterprise risk management task but computationally challenging due to the need for simulation. Existing methods in literature only use simple experimental designs with significant room improvement. This article identifies three major components efficient framework. In addition, we propose optimal and provides analytical insights each component. Our numerical results show that our proposal achieves significantly higher...
In recent years, Convolution Neural Network (CNN) has achieved a series of breakthrough results in large-scale identification tasks such as image classification and recognition, target location detection. There are severe speckle noise distortion Synthetic Aperture Radar (SAR) image, the traditional method low recognition rate robustness to noise. To solve these problems, novel approach for noisy SAR based on fusion filter CNN is suggested. Firstly, suppression carried out by using filtering...