Zhiyuan Zhang

ORCID: 0000-0003-0124-9791
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Adversarial Robustness in Machine Learning
  • Electrocatalysts for Energy Conversion
  • Natural product bioactivities and synthesis
  • Catalytic Processes in Materials Science
  • Natural Language Processing Techniques
  • Financial Markets and Investment Strategies
  • Stock Market Forecasting Methods
  • Data Stream Mining Techniques
  • Text and Document Classification Technologies
  • Advanced Photocatalysis Techniques
  • Machine Learning and ELM
  • Circular RNAs in diseases
  • CO2 Reduction Techniques and Catalysts
  • Carbon dioxide utilization in catalysis
  • Ginseng Biological Effects and Applications
  • Embedded Systems and FPGA Design
  • MicroRNA in disease regulation
  • Transplantation: Methods and Outcomes
  • Helminth infection and control
  • Explainable Artificial Intelligence (XAI)
  • Ammonia Synthesis and Nitrogen Reduction
  • Neural Networks and Applications
  • Pharmacogenetics and Drug Metabolism
  • Salivary Gland Disorders and Functions
  • Osteoarthritis Treatment and Mechanisms

Peking University
2008-2024

Tsinghua–Berkeley Shenzhen Institute
2021-2024

Tsinghua University
2021-2024

Nanjing Institute of Technology
2024

Institute for Learning Innovation
2024

Shanghai Pulmonary Hospital
2024

Tongji University
2024

Foshan University
2023

Tianjin University of Traditional Chinese Medicine
2022

Tianjin International Joint Academy of Biomedicine
2020

Layer normalization (LayerNorm) is a technique to normalize the distributions of intermediate layers. It enables smoother gradients, faster training, and better generalization accuracy. However, it still unclear where effectiveness stems from. In this paper, our main contribution take step further in understanding LayerNorm. Many previous studies believe that success LayerNorm comes from forward normalization. Unlike them, we find derivatives mean variance are more important than by...

10.48550/arxiv.1911.07013 preprint EN other-oa arXiv (Cornell University) 2019-01-01

Landscape wildfires generate a substantial amount of dissolved black carbon (DBC) annually, yet the molecular nitrogen (N) structures in DBC are poorly understood. Here, we systematically compared chemodiversity N-containing molecules among three different samples from rice straw biochar pyrolyzed at 300, 400, and 500 °C, one leached organic (LDOC) sample composted straw, fire-affected soil matter (SDOMFire) using Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR MS)....

10.1021/acs.est.4c01829 article EN Environmental Science & Technology 2024-06-27

Circular RNAs (circRNAs), a novel class of endogenous long non-coding RNAs, have attracted considerable attention due to their closed continuous loop structure and potential clinical value. In this study, we investigated the function circFASTKD1 in vascular endothelial cells. CircFASTKD1 bound directly miR-106a relieved its inhibition Large Tumor Suppressor Kinases 1 2, thereby suppressing Yes-Associated Protein signaling pathway. Under both normal hypoxic conditions, ectopic expression...

10.18632/aging.202305 article EN cc-by Aging 2020-12-19

Albendazole is the preferred deworming drug and has strong insecticidal effects on human animal helminth parasites, showing remarkable activity against hepatocellular carcinoma colorectal cancer cells. However, it classified as being in class II Biopharmaceutics Classification System due to its poor water solubility (0.2 mg/L) high permeability, which make clinical application of albendazole impractical. Through complexation with methyl-β-cyclodextrin, best result so far, albendazole's was...

10.3390/molecules28217295 article EN cc-by Molecules 2023-10-27

We argue that the vulnerability of model parameters is crucial value to study robustness and generalization but little research has been devoted understanding this matter. In work, we propose an indicator measure neural network by exploiting their via parameter corruption. The proposed describes maximum loss variation in non-trivial worst-case scenario under For practical purposes, give a gradient-based estimation, which far more effective than random corruption trials can hardly induce...

10.1609/aaai.v35i13.17385 article EN Proceedings of the AAAI Conference on Artificial Intelligence 2021-05-18

The electrochemical reduction of CO2 to CH4 is promising for carbon neutrality and renewable energy storage but confronts low selectivity, especially at high current densities. key challenge lies in promoting *CO intermediate *H coupling while minimizing side reactions including C‐C or H‐H coupling, which particularly difficult density due abundant intermediates. Here we report a cooperative strategy address this using Cu‐based catalysts comprising Cu‐N coordination polymer CuO component...

10.1002/anie.202423915 article EN Angewandte Chemie International Edition 2024-12-26

Methyl 3-amino-6-methoxythieno [2,3-b] quinoline-2-carboxylate (PU-48) is a novel diuretic urea transporter inhibitor. The aim of this study to investigate the profile plasma pharmacokinetics, tissue distribution, and excretion by oral dosing PU-48 in rats. Concentrations within biological samples are determined using validated high performance liquid chromatography-tandem mass spectrometry (LC-MS/MS) method. After administration (3, 6, 12 mg/kg, respectively) self-nanomicroemulsifying drug...

10.3390/pharmaceutics10030124 article EN cc-by Pharmaceutics 2018-08-08

Recent studies have revealed a security threat to natural language processing (NLP) models, called the Backdoor Attack. Victim models can maintain competitive performance on clean samples while behaving abnormally with specific trigger word inserted. Previous backdoor attacking methods usually assume that attackers certain degree of data knowledge, either dataset which users would use or proxy datasets for similar task, implementing poisoning procedure. However, in this paper, we find it is...

10.48550/arxiv.2103.15543 preprint EN cc-by arXiv (Cornell University) 2021-01-01

A systematic study on pharmacokinetcs and tissue distribution of danshensu (one the major active components from Salvia miltiorrhiza) is conducted using a rapid sensitive highperformance liquid chromatographic (HPLC) method. Before HPLC analysis, biological samples are pretreated with liquid-liquid extraction. Separation internal standard achieved an Agilent Zorbax C18 column mobile phase made up acetonitrile 0.05% trifluoracetic acid at flow rate 0.8 mL/min. The calibration curves in plasma...

10.1093/chromsci/46.2.184 article EN Journal of Chromatographic Science 2008-02-01

A specific, sensitive and stable high-performance liquid chromatographic-tandem mass spectrometry (LC-MS/MS) method was developed validated for the quantitative determination of methyl 3-amino-6-methoxythieno [2,3-b]quinoline-2-carboxylate (PU-48), a novel diuretic thienoquinolin urea transporter inhibitor in rat plasma. In this method, chromatographic separation PU-48 achieved with reversed-phase C18 column (100 × 2.1 mm, 3 μm) at 35°C. The mobile phase consisted acetonitrile water 0.05%...

10.1002/bmc.4157 article EN Biomedical Chromatography 2017-11-29

Abstract The imbalance of steroid hormones is closely related to the occurrence and development hepatocellular carcinoma (HCC). However, most research has focused on hormone receptors, reports about relationship between serum concentration cortisol HCC are rare. aim this was establish a simple, specific, sensitive reliable liquid chromatography–mass spectrometry/mass spectrometry (LC–MS/MS) method for quantitation in human compare level 221 patients 183 healthy volunteers. results showed...

10.1002/bmc.5829 article EN Biomedical Chromatography 2024-02-13

Motivation: Deep learning methods have achieved superior reconstruction in MR quantitative T1ρ imaging due to their ability learn the non-linearity relationship between undersampled k-space data and corresponding maps. Goal(s): In this study, we investigate use of DDPM for highly accelerated imaging. Approach: The learns image properties from fully acquired images during training without knowledge subsampling patterns used scans. This is advantageous most existing models that need be...

10.58530/2024/4555 article EN Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition 2024-11-26

Electrocatalytic organic oxidation coupled with hydrogen production emerges as a profitable solution to simultaneously reduce overall energy consumption of H2 and synthetic high‐value chemicals. Noble metal catalysts are highly efficient electrocatalysts in reactions, but they deactivate easily weakening the benefit actual production. Herein, we report universal asymmetric pulse potential strategy achieve long‐term stable operation noble metals for various alcohol reactions catalysts. For...

10.1002/ange.202420992 article EN Angewandte Chemie 2024-12-08

Electrocatalytic organic oxidation coupled with hydrogen production emerges as a profitable solution to simultaneously reduce overall energy consumption of H2 and synthetic high‐value chemicals. Noble metal catalysts are highly efficient electrocatalysts in reactions, but they deactivate easily weakening the benefit actual production. Herein, we report universal asymmetric pulse potential strategy achieve long‐term stable operation noble metals for various alcohol reactions catalysts. For...

10.1002/anie.202420992 article EN Angewandte Chemie International Edition 2024-12-08

Mizolastine is a selective histamine H1 receptor antagonist for chronic urticaria or allergic rhinitis. We investigated whether the variant genotypes of metabolic enzymes UGT1A1, CYP3A5 and transporter ABCB1 influence pharmacokinetic phenotype substrate mizolastine in Chinese volunteers. Genotyping single nucleotide polymorphisms UGT1A1*6 (G211A), CYP3A5*3 (A6986G) (C3435T) was determined by pyrosequencing method. After oral dose 10 mg mizolastine, plasma concentrations were measured using...

10.1111/bcpt.13028 article EN Basic & Clinical Pharmacology & Toxicology 2018-04-27
Coming Soon ...