Search

Your search keyword '"Ma Fang"' showing total 3,672 results
Start Over Author "Ma Fang"
3,672 results on '"Ma Fang"'

64. Exploiting Position Bias for Robust Aspect Sentiment Classification

Author

Ma, Fang, Zhang, Chen, and Song, Dawei

Subjects
Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

Aspect sentiment classification (ASC) aims at determining sentiments expressed towards different aspects in a sentence. While state-of-the-art ASC models have achieved remarkable performance, they are recently shown to suffer from the issue of robustness. Particularly in two common scenarios: when domains of test and training data are different (out-of-domain scenario) or test data is adversarially perturbed (adversarial scenario), ASC models may attend to irrelevant words and neglect opinion expressions that truly describe diverse aspects. To tackle the challenge, in this paper, we hypothesize that position bias (i.e., the words closer to a concerning aspect would carry a higher degree of importance) is crucial for building more robust ASC models by reducing the probability of mis-attending. Accordingly, we propose two mechanisms for capturing position bias, namely position-biased weight and position-biased dropout, which can be flexibly injected into existing models to enhance representations for classification. Experiments conducted on out-of-domain and adversarial datasets demonstrate that our proposed approaches largely improve the robustness and effectiveness of current models., Comment: 7 pages, 2 figures, 4 tables, accepted to Findings of ACL 2021. Repo: https://github.com/BD-MF/POS4ASC

Published
2021

74. A dynamic plasma membrane proteome analysis of alcohol-induced liver cirrhosis

Author

Jia Xiaofang, Yin Lin, Feng Yanling, Peng Xia, Ma Fang, Yao Yamin, Liu Xiaoqian, Zhang Zhiyong, Yuan Zhenghong, and Zhang Lijun

Subjects
Alcohol, Liver cirrhosis, Plasma membrane, Proteomics, Cytology, QH573-671
Abstract

Abstract Alcohol-induced injury has become one of the major causes for liver cirrhosis. However, the molecular mechanisms of ethanol-induced injury are not fully understood. To this end, we performed a dynamic plasma membrane proteomic research on rat model. A rat model from hepatitis to liver cirrhosis was developed. Plasma membrane from liver tissue with liver fibrosis stage of 2 and 4 (S2 and S4) was purified by sucrose density gradient centrifugation. Its purification was verified by western blotting. Proteins from plasma membrane were separated by two-dimensional electrophoresis (2DE) and differentially expressed proteins were identified by tandem mass spectrometry. 16 consistent differentially expressed proteins from S2 to S4 were identified by mass spectrometry. The expression of differentially expressed proteins annexin A6 and annexin A3 were verified by western blotting, and annexin A3 was futher verified by immunohistochemistry. Our research suggests a possible mechanism by which ethanol alters protein expression to enhance the liver fibrosis progression. These differentially expressed proteins might be new drug targets for treating alcoholic liver cirrhosis.

Published
2012
Full Text
View/download PDF

75. Comparative proteomic analysis of plasma membrane proteins between human osteosarcoma and normal osteoblastic cell lines

Author

Peng Xia, Hu Shuo, Li Jian, Jia Xiaofang, Hua Yingqi, Zhang Lijun, Zhang Zhiyu, Yang Pengyuan, Sun Mengxiong, Ma Fang, and Cai Zhengdong

Subjects
Neoplasms. Tumors. Oncology. Including cancer and carcinogens, RC254-282
Abstract

Abstract Background Osteosarcoma (OS) is the most common primary malignant tumor of bone in children and adolescents. However, the knowledge in diagnostic modalities has progressed less. To identify new biomarkers for the early diagnosis of OS as well as for potential novel therapeutic candidates, we performed a sub-cellular comparative proteomic research. Methods An osteosarcoma cell line (MG-63) and human osteoblastic cells (hFOB1.19) were used as our comparative model. Plasma membrane (PM) was obtained by aqueous two-phase partition. Proteins were analyzed through iTRAQ-based quantitative differential LC/MS/MS. The location and function of differential proteins were analyzed through GO database. Protein-protein interaction was examined through String software. One of differentially expressed proteins was verified by immunohistochemistry. Results 342 non-redundant proteins were identified, 68 of which were differentially expressed with 1.5-fold difference, with 25 up-regulated and 43 down-regulated. Among those differential proteins, 69% ware plasma membrane, which are related to the biological processes of binding, cell structure, signal transduction, cell adhesion, etc., and interaction with each other. One protein--CD151 located in net nodes was verified to be over-expressed in osteosarcoma tissue by immunohistochemistry. Conclusion It is the first time to use plasma membrane proteomics for studying the OS membrane proteins according to our knowledge. We generated preliminary but comprehensive data about membrane protein of osteosarcoma. Among these, CD151 was further validated in patient samples, and this small molecule membrane might be a new target for OS research. The plasma membrane proteins identified in this study may provide new insight into osteosarcoma biology and potential diagnostic and therapeutic biomarkers.

Published
2010
Full Text
View/download PDF

76. Proteomic analysis of PBMCs: characterization of potential HIV-associated proteins

Author

Yin Lin, Ma Fang, Qi Tangkai, Peng Xia, Sun Jianjun, Zhang Xiaojun, Jia Xiaofang, Zhang Lijun, Yao Yamin, Qiu Chao, and Lu Hongzhou

Subjects
Cytology, QH573-671
Abstract

Abstract Background The human immunodeficiency virus type 1 (HIV-1) pandemic has continued unabated for nearly 30 years. To better understand the influence of virus on host cells, we performed the differential proteome research of peripheral blood mononuclear cells (PBMCs) from HIV-positive patients and healthy controls. Results 26 protein spots with more than 1.5-fold difference were detected in two dimensional electrophoresis (2DE) gels. 12 unique up-regulated and one down-regulated proteins were identified in HIV-positive patients compared with healthy donors. The mRNA expression of 10 genes was analyzed by real time RT-PCR. It shows that the mRNA expression of talin-1, vinculin and coronin-1C were up-regulated in HIV positive patients and consistent with protein expression. Western blotting analysis confirmed the induction of fragments of vinculin, talin-1 and filamin-A in pooled and most part of individual HIV-positive clinical samples. Bioinformatic analysis showed that a wide host protein network was disrupted in HIV-positive patients. Conclusions Together, this work provided useful information to facilitate further investigation of the underlying mechanism of HIV and host cell protein interactions, and discovered novel potential biomarkers such as fragment of vinculin, filamin-A and talin-1 for anti-HIV research.

Published
2010
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results