Search

Your search keyword '"Chen, Shaoyi"' showing total 135 results
Start Over Author "Chen, Shaoyi" Search Limiters Full Text Publication Year Range Last 50 years
135 results on '"Chen, Shaoyi"'

4. Text2Event: Controllable Sequence-to-Structure Generation for End-to-end Event Extraction

Author

Lu, Yaojie, Lin, Hongyu, Xu, Jin, Han, Xianpei, Tang, Jialong, Li, Annan, Sun, Le, Liao, Meng, and Chen, Shaoyi

Subjects
Computer Science - Computation and Language
Abstract

Event extraction is challenging due to the complex structure of event records and the semantic gap between text and event. Traditional methods usually extract event records by decomposing the complex structure prediction task into multiple subtasks. In this paper, we propose Text2Event, a sequence-to-structure generation paradigm that can directly extract events from the text in an end-to-end manner. Specifically, we design a sequence-to-structure network for unified event extraction, a constrained decoding algorithm for event knowledge injection during inference, and a curriculum learning algorithm for efficient model learning. Experimental results show that, by uniformly modeling all tasks in a single model and universally predicting different labels, our method can achieve competitive performance using only record-level annotations in both supervised learning and transfer learning settings., Comment: Accepted to ACL2021 (main conference)

Published
2021

5. Comparative Studies on Microbial Community Structure and Production Performance of Jiang-Flavor Daqu in Different Areas of Maotai Town

Author

CHEN Shaoyi, LANG Ying, QIU Shuyi, WU Botian, HU Shenglan, ZHOU Hongxiang

Subjects
microbial community structure, physicochemical indexes, characteristic flavor substances, correlation analysis, redundancy analysis, Food processing and manufacture, TP368-456
Abstract

The microbial community structure and diversity of Jiang-flavor Daqu (TS, WS, WM, MH and DJ) from different areas of Maotai town were analyzed by using the third-generation nanopore sequencing platform, and its physicochemical indexes and characteristic flavor substances were measured. The results showed that there were some similarities and differences between Daqu in different areas of Maotai town. In terms of microbial community structure, Bacillus, Saccharopolyspora, Weissella, Staphylococcus and Streptomyces were the common dominant bacterial genera in the five Daqu samples. Among them, Bacillus was the absolute dominant bacteria in MH and DJ. Aspergillus and Penicillium were the common dominant fungal genera, and the proportions of Lichtheimia and Saccharomycopsis in TS, WM and MH were significantly higher than those in DJ and WS. Compared with TS and WM, network correlation analysis showed that MH, DJ and WS had stronger interactions among dominant bacteria. In addition, redundancy analysis (RDA) showed that Weissella was positively correlated with esterification power, liquefaction power, saccharification power, acetic acid, ethyl acetate, ethyl lactate and n-pentanol. Lichtheimia was positively correlated with liquefaction power, saccharification power, acetic acid, isovaleric acid, 2,3-butanediol, phenylacetaldehyde and dibutyl phthalate. Saccharomycopsis was positively correlated with esterification power and ethyl acetate. Bacillus was positively correlated with 2,3,5,6-tetramethylpyrazine, propionic acid, isovaleric acid, dibutyl phthalate, 2,3-butanediol and phenacetaldehyde.

Published
2023
Full Text
View/download PDF

18. SHH/GLI2-TGF-β1 feedback loop between cancer cells and tumor-associated macrophages maintains epithelial-mesenchymal transition and endoplasmic reticulum homeostasis in cholangiocarcinoma

Author

Chen, Zuxiao, primary, Li, Haiyan, additional, Li, Zongyan, additional, Chen, Shaoyi, additional, Huang, Xiaoming, additional, Zheng, Zheyu, additional, Qian, Xiangjun, additional, Zhang, Lei, additional, Long, Guojie, additional, Xie, Jiancong, additional, Wang, Qian, additional, Pan, Weidong, additional, and Zhang, Dawei, additional

Published
2023
Full Text
View/download PDF

21. Mesenchymal Stem Cell-derived Extracellular Vesicles Prevent Experimental Bronchopulmonary Dysplasia Complicated By Pulmonary Hypertension

Author

Sharma, Mayank, primary, Bellio, Michael A, additional, Benny, Merline, additional, Kulandavelu, Shathiyah, additional, Chen, Pingping, additional, Janjindamai, Chawisa, additional, Han, Chenxu, additional, Chang, Liming, additional, Sterling, Shanique, additional, Williams, Kevin, additional, Damianos, Andreas, additional, Batlahally, Sunil, additional, Kelly, Kaitlyn, additional, Aguilar-Caballero, Daniela, additional, Zambrano, Ronald, additional, Chen, Shaoyi, additional, Huang, Jian, additional, Wu, Shu, additional, Hare, Joshua M, additional, Schmidt, Augusto, additional, Khan, Aisha, additional, and Young, Karen, additional

Published
2022
Full Text
View/download PDF

22. Active Learning with Query Generation for Cost-Effective Text Classification

Author

Jin Xu, Meng Liao, Sheng-Jun Huang, Chen Shaoyi, and Yi Fan Yan

Subjects
0209 industrial biotechnology, Active learning (machine learning), business.industry, Computer science, Scale (chemistry), media_common.quotation_subject, 02 engineering and technology, General Medicine, Machine learning, computer.software_genre, Class (biology), Annotation, 020901 industrial engineering & automation, Reading (process), Active learning, 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, Relevance (information retrieval), Artificial intelligence, business, computer, media_common
Abstract

Labeling a text document is usually time consuming because it requires the annotator to read the whole document and check its relevance with each possible class label. It thus becomes rather expensive to train an effective model for text classification when it involves a large dataset of long documents. In this paper, we propose an active learning approach for text classification with lower annotation cost. Instead of scanning all the examples in the unlabeled data pool to select the best one for query, the proposed method automatically generates the most informative examples based on the classification model, and thus can be applied to tasks with large scale or even infinite unlabeled data. Furthermore, we propose to approximate the generated example with a few summary words by sparse reconstruction, which allows the annotators to easily assign the class label by reading a few words rather than the long document. Experiments on different datasets demonstrate that the proposed approach can effectively improve the classification performance while significantly reduce the annotation cost.

Published
2020

23. Author Correction: Hyperoxia-activated circulating extracellular vesicles induce lung and brain injury in neonatal rats

Author

Ali, Anum, primary, Zambrano, Ronald, additional, Duncan, Matthew R., additional, Chen, Shaoyi, additional, Luo, Shihua, additional, Yuan, Huijun, additional, Chen, Pingping, additional, Benny, Merline, additional, Schmidt, Augusto, additional, Young, Karen, additional, Kerr, Nadine, additional, de Rivero Vaccari, Juan Pablo, additional, Keane, Robert W., additional, Dietrich, W. Dalton, additional, and Wu, Shu, additional

Published
2021
Full Text
View/download PDF

24. Text2Event: Controllable Sequence-to-Structure Generation for End-to-end Event Extraction

Author

Meng Liao, Chen Shaoyi, Jin Xu, Le Sun, Xianpei Han, Hongyu Lin, Annan Li, Yaojie Lu, and Jialong Tang

Subjects
Structure (mathematical logic), FOS: Computer and information sciences, Computer Science - Computation and Language, business.industry, Event (computing), Computer science, Supervised learning, Inference, Machine learning, computer.software_genre, Task (project management), Artificial intelligence, business, Transfer of learning, computer, Computation and Language (cs.CL), Decoding methods, Semantic gap
Abstract

Event extraction is challenging due to the complex structure of event records and the semantic gap between text and event. Traditional methods usually extract event records by decomposing the complex structure prediction task into multiple subtasks. In this paper, we propose Text2Event, a sequence-to-structure generation paradigm that can directly extract events from the text in an end-to-end manner. Specifically, we design a sequence-to-structure network for unified event extraction, a constrained decoding algorithm for event knowledge injection during inference, and a curriculum learning algorithm for efficient model learning. Experimental results show that, by uniformly modeling all tasks in a single model and universally predicting different labels, our method can achieve competitive performance using only record-level annotations in both supervised learning and transfer learning settings., Accepted to ACL2021 (main conference)

Published
2021

26. Field Trials of Communication and Sensing System in Space Division Multiplexing Optical Fiber Cable

Author

Chen, Yingyu, Xiao, Yongguang, Chen, Shaoyi, Xie, Fengmao, Liu, Zhengyong, Wang, Dawei, Yi, Xingwen, Lu, Chao, and Li, Zhaohui

Abstract

Space division multiplexing (SDM) technology has become a research hotspot, which can increase the transmission capacity extensively and is supposed to be the next generation of the optical communication technology. Although it has been demonstrated technically in the laboratory, SDM based on multicore optical fiber still lacks of convinced field trails as demonstration and verification to commercialize such technology. In this work as a field test, we deployed a seven-core optical fiber cable for application in the practical scenarios and successfully demonstrated large capacity SDM communication system, reaching a capacity of 234 Tb/s (net 187.49 Tb/s) and achieving the capacity-distance product of 61.87 Pb/s·km (net 49.50 Pbit/s·km) in one strand of fiber just in C band. To our knowledge, it is the largest capacity in the world based on the deployed SDM multicore optical fiber cable. Additionally, we further showed the utilization in sensing that monitors the traffic surrounding the deployed seven-core fiber cable. It is knowledgeable that SDM technology requires plenty of evident verifications to overcome technical and engineering issues prior to practical deployment, and the successful field test with promising results shows its great potential of employing SDM technology in a demand of large capacity and smart city, moving it toward commercialization.

Published
2023
Full Text
View/download PDF

28. Additional file 1 of Circulating extracellular vesicles activate the pyroptosis pathway in the brain following ventilation-induced lung injury

Author

Chavez, Laura, Meguro, Julia, Chen, Shaoyi, de Paiva, Vanessa Nunes, Zambrano, Ronald, Eterno, Julia M., Kumar, Rahul, Duncan, Matthew R., Benny, Merline, Young, Karen C., Dietrich, W. Dalton, Brambilla, Roberta, Wu, Shu, and Schmidt, Augusto F.

Abstract

Additional file 1: Table S1. Primary antibodies��� information and dilutions used. Table S2. Secondary antibodies��� information and dilutions used. Table S3. Multiplex ELISA of lung, plasma, and corpus callosum of neonatal rats.

Published
2021
Full Text
View/download PDF

40. Interaction and biodegradation evaluate of m -cresol and quinoline in co-exist system

Author

Guobao Zhao, Chen Shaoyi, Yuan Ren, and Chaohai Wei

Subjects
m-Cresol, Lysinibacillus, Achromobacter, biology, Quinoline, Biodegradation, biology.organism_classification, Microbiology, Biomaterials, chemistry.chemical_compound, chemistry, Wastewater, Degradation (geology), Waste Management and Disposal, Bacteria, Nuclear chemistry
Abstract

Environmentally hazardous and toxic chemicals are commonly generated in actual wastewater that the complex compositions in wastewater treatment system need different types of strains to be degraded. The main objective of this research was to understand the effect of extra substrates, phenolic and N-heterocyclic compounds, on the performance of pure cultural and mixed strains under single and dual substrates conditions. Two bacteria, Lysinibacillus sp. SC03 and Achromobacter sp. DN-06, were acclimated to degrade different concentrations of m-cresol and quinoline. The results indicated that Lysinibacillus sp. SC03 could completely degrade 100 mg l−1 m-cresol with no delay time, however, little removal of quinoline was observed; Achromobacter sp. DN-06 could degrade 100 mg l−1 quinoline in 32 h, but could not remove m-cresol, which means m-cresol and quinoline is the specific substrate. The degradation rate of m-cresol fitted well to the zero-order kinetic equation although the degrading ability of Lysinibacillus sp. SC03 was inhibited when less than 100 mg l−1 quinoline was added, and the inhibitive effect was confirmed to be a noncompetitive pattern which could be interpreted by the Michaelis–Menten kinetics equation with corresponding parameters Vmax, Km, K1 and K2 were 13.16 mg l−1 h−1, 35.84 mg l−1, 200.0 mg l−1 and 285.7 mg l−1, respectively. Moreover, the addition of m-cresol-degrading strain (Lysinibacillus sp. SC03) could accelerate the removal of quinoline because the metabolites of quinoline could be degraded by Lysinibacillus sp. SC03 and the chemical equilibrium moved to more biodegradation of quinoline. Also, this process attributed to less the delay time during the quinoline removal.

Published
2014

41. Nitrogen‐Doped Hard Carbon on Nickel Foam as Free‐Standing Anodes for High‐Performance Sodium‐Ion Batteries.

Author

Li, Ruizi, Huang, Jianfeng, Li, Jiayin, Cao, Liyun, Luo, Yijia, He, Yuanyuan, Lu, Guoxing, Yu, Aimin, and Chen, Shaoyi

Subjects
ANODES, GRAPHITIZATION, ELECTRIC batteries, CHEMICAL stability, NITROGEN, CARBON foams, ELECTRON transport
Abstract

Nitrogen‐doped hard carbon on Ni foam (NC/NF) with numerous active sites, expanded interlayer distance (0.49 nm), and highly ordered pseudo‐graphic structure is synthesized by a melamine‐assisted hydrothermal pretreatment and subsequent pyrolysis procedure. The high content configurations of pyrrolic‐nitrogen and pyridinic‐nitrogen in NC/NF could provide abundant active sites and defects for efficient Na+ adsorption/desorption and improved transport kinetics. The obvious flocculent structures on carbon surface, corresponding to the highly ordered pseudo‐graphic structure, could enhance the transport kinetics for electrons. The strong electrostatic repulsion of pyrrolic‐nitrogen could expand interlayer distance, facilitating Na+ insertion/extraction process. The balance between high nitrogen content and sufficient graphitization degree comes into strong carbon frameworks with enhanced structural stability and electronic conductivity. When served as free‐standing anodes for sodium‐ion batteries (SIBs), NC/NF presents high rate capability of 128.5 mAh g−1 at 10 A g−1, and excellent cycling stability of 225.4 mAh g−1 after 1000 cycles under high current density of 5 A g−1. These results demonstrate a new light to prepare advanced anodes for SIBs. [ABSTRACT FROM AUTHOR]

Published
2020
Full Text
View/download PDF

43. Influence of tungsten oxides’ characteristics on fineness, homogeneity and looseness of reduced ultrafine tungsten powder

Author

Zou Zhi-qiang, Huang Bai-yun, Chen Shaoyi, Lu Haibo, and Liao Jiqiao

Subjects
Materials science, chemistry, Hydrogen, Powder metallurgy, Fineness, Homogeneity (physics), Metallurgy, Chemical reduction, chemistry.chemical_element, Metal powder, Tungsten
Abstract

This paper deals with the influence of the characteristics of tungsten oxides on fineness, homogeneity and looseness of tungsten powder by conventional hydrogen reduction techniques.

Published
1999

Catalog

Books, media, physical & digital resources
See catalog results