Chunxi Liu

Qilu Hospital of Shandong University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Na Zhang

Huaqiao University

Yatharth Saraf

Meta (Israel)

Sanjeev Khudanpur

Johns Hopkins University

Geoffrey Zweig

Meta (Israel)

Qingming Huang

University of Chinese Academy of Sciences

Najim Dehak

Johns Hopkins University

Yingjie Zhang

Shandong University

Xiaoxiao Qian

Nanjing Tech University

Frank Zhang

Yale University

Shuqiang Jiang

Chinese Academy of Sciences

Cooperative Institutions

Shandong University

105

Chinese Academy of Sciences

Qilu Hospital of Shandong University

Chinese Academy of Medical Sciences & Peking Union Medical College

University of Chinese Academy of Sciences

Peking University

Shanghai Jiao Tong University

Meta (Israel)

Zhejiang University

Meta (United States)

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

[Expressions of cyclin D1 and p16 in hypopharyngeal squamous cell carcinoma and their clinical significance].

PubMed (2007)

Zhijie Fu Xinliang Pan Dapeng Lei Dayu Liu Wenjun Liu

To study the expressions of Cyclin D1 and p16 proteins in hypopharyngeal squamous cell carcinoma and their clinical significance.Immunohistochemical technology (P-V) was applied to detect the expression of Cyclin D1 and p16 in 36 cases of hypopharyngeal squamous cell carcinoma and 10 cases of normal epithelium.(1) The expression of cyclin D1 in the tumorous cell was significantly higher than that in normal epithelium (P < 0.05). The expression of p16 in the tumorous cell was significantly lower than that in normal epithelium (P < 0.05); (2) The positive rate of Cyclin D1 was significantly correlated with clinical stage and cervical lymph node metastases (P < 0.05). The positive rate of p16 was correlated with cervical lymph node metastases (P < 0.05). The positive rates of Cyclin D1 and p16 were not correlated with the gender, age, tumor size, primary lesion and cell differentiation (P > 0.05); (3) There was correlation between the expression of Cyclin D1 and the expression of p16 (r(s) = -0.420, P < 0.05).The over-expression of Cyclin D1 and the under-expression of p16 may play a significant role in the occurrence incidence and development of hypopharyngeal squamous cell carcinoma, and may be important indicators for cervical lymph node metastases.

Cyclin D

Cyclin B

Source

Cite

Citations (1)

Improving RNN Transducer Based ASR with Auxiliary Tasks

2022 IEEE Spoken Language Technology Workshop (SLT) (2021)

Chunxi Liu Feida Zhang Duc Le Suyoun Kim Yatharth Saraf

End-to-end automatic speech recognition (ASR) models with a single neural network have recently demonstrated state-of-the-art results compared to conventional hybrid speech recognizers. Specifically, recurrent neural network transducer (RNN-T) has shown competitive ASR performance on various benchmarks. In this work, we examine ways in which RNN-T can achieve better ASR accuracy via performing auxiliary tasks. We propose (i) using the same auxiliary task as primary RNN-T ASR task, and (ii) performing context-dependent graphemic state prediction as in conventional hybrid modeling. In transcribing social media videos with varying training data size, we first evaluate the streaming ASR performance on three languages: Romanian, Turkish and German. We find that both proposed methods provide consistent improvements. Next, we observe that both auxiliary tasks demonstrate efficacy in learning deep transformer encoders for RNN-T criterion, thus achieving competitive results -2.0%/4.2% WER on LibriSpeech test-clean/other - as compared to prior top performing models.

10.1109/slt48900.2021.9383548

Cite

Citations (33)

Streaming Transformer Transducer based Speech Recognition Using Non-Causal Convolution

ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2022)

Yangyang Shi Chunyang Wu Dilin Wang Alex Xiao Jay Mahadeokar

This paper improves the streaming transformer transducer for speech recognition using non-causal convolution. Many works apply the causal convolution to improve streaming transformer ignoring the lookahead context. We propose to use non-causal convolution to process the center block and lookahead context separately. This method leverages the lookahead context in convolution and maintains similar training and decoding efficiency. Given the similar latency, using the non-causal convolution with lookahead context gives better accuracy than causal convolution, especially for open-domain dictation. Besides, this paper applies talking-head attention and a novel history context compression scheme to further improve the performance. The talking-head attention improves the multi-head self-attention by transferring information among different heads. The history context compression method introduces more extended history context compactly. On our in-house data, the proposed methods improve a small Emformer baseline with lookahead context by relative WERR 5.1%, 14.5%, 8.4% on open-domain dictation, assistant general scenarios, and assistant calling scenarios respectively.

Convolution (computer science)

Dictation

Viterbi decoder

Overlap–add method

10.1109/icassp43922.2022.9747706

Cite

Citations (10)

Enhanced gene transfection efficiency in CD13-positive vascular endothelial cells with targeted poly(lactic acid)–poly(ethylene glycol) nanoparticles through caveolae-mediated endocytosis

Journal of Controlled Release (2011)

Chunxi Liu Wangyang Yu Zhijin Chen Jian Zhang Na Zhang

Internalization

10.1016/j.jconrel.2011.02.027

Cite

Citations (60)

Synthesis, biological and pharmacokinetic characterization of a novel leucine ureido derivative as a multi-target anticancer agent

RSC Advances (2025)

Fangyuan Shi Qifu Xu Yingjie Zhang Jiangying Cao Chunxi Liu

A new leucine ureido derivative compound 3 exhibited potent antitumor properties, excellent in vitro stability and in vivo pharmacokinetic properties, was a promising lead for further development.

Derivative (finance)

Lead compound

10.1039/d4ra03200d

Cite

Citations (0)

Pretraining by Backtranslation for End-to-End ASR in Low-Resource Settings

Interspeech 2022 (2019)

Matthew Wiesner Adithya Renduchintala Shinji Watanabe Chunxi Liu Najim Dehak

We explore training attention-based encoder-decoder ASR in low-resource settings. These models perform poorly when trained on small amounts of transcribed speech, in part because they depend on having sufficient target-side text to train the attention and decoder networks. In this paper we address this shortcoming by pretraining our network parameters using only text-based data and transcribed speech from other languages. We analyze the relative contributions of both sources of data. Across 3 test languages, our text-based approach resulted in a 20% average relative improvement over a text-based augmentation technique without pretraining. Using transcribed speech from nearby languages gives a further 20-30% relative reduction in character error rate.

End-to-end principle

10.21437/interspeech.2019-3254

Cite

Citations (11)

Neurotoxicity of p-amyloid in Alzheimer' s disease

Chunxi Liu

Alzheimer's disease (AD) is the most common form of dementia among the elderly. However, the etiology and pathogenic mechanisms of AD are not fully understood. Recent evidence suggests that β-amyloid protein (Aβ) plays a key role in the development of Alzheimer's disease. This review focuses on the neurotoxicity mechanisms of Aβ.

Neurotoxicity

Amyloid (mycology)

Etiology

Cite

Citations (0)

Discovery of a Novel Hybrid of Vorinostat and Riluzole as a Potent Antitumor Agent

Frontiers in Cell and Developmental Biology (2020)

Qifu Xu Chunxi Liu Jie Zang Shuai Gao C. James Chou

Vorinostat (suberoylanilide hydroxamic acid) was the first approved histone deacetylase (HDAC) inhibitor in a group of validated cancer therapeutic agents targeting epigenetics. Riluzole is a drug used to treat amyotrophic lateral sclerosis, the antitumor potency of which has been recently revealed. Herein, a novel hybrid of vorinostat and riluzole (compound 1) was rationally designed, synthesized, and evaluated. Compared with vorinostat, compound 1 exhibited superior total HDAC inhibitory activity and similar HDAC isoform selective profiles. The intracellular HDAC inhibition of compound 1 was confirmed by Western blot analysis. Moreover, compound 1 possessed more potent in vitro antiproliferative activity against all tested solid and hematological tumor cell lines than vorinostat. In vitro metabolic stability evaluation of compound 1 revealed better human plasma stability and comparable human liver microsomal stability than vorinostat. Additionally, compound 1 demonstrated more significant in vivo antitumor activity in a MDA-MB-231 xenograft model than vorinostat, which could be attributed to its superior in vitro antiproliferative activity and metabolic stability. Taken together, the results presented here support further research and development of compound 1 as a promising antitumor agent.

Riluzole

Vorinostat

10.3389/fcell.2020.00454

Cite

Citations (14)

Perovskite cesium lead bromide quantum dots: A new efficient photocatalyst for degrading antibiotic residues in organic system

Journal of Cleaner Production (2019)

Xiaoxiao Qian Zhi Chen Xiuru Yang Wan Zhao Chunxi Liu

Photodegradation

Methyl orange

Silver bromide

10.1016/j.jclepro.2019.119335

Cite

Citations (86)

Reform of the three-stage experimental teaching of medical laboratory technology

Zhonghua yixue jiaoyu tansuo zazhi (2016)

Chunyan Gao Shuting Ji Xinying Zhang Yingli Chen Chunxi Liu

In order to adapt to the training target of medical laboratory technology professionals under the condition of five to four educational system, we reformed the experimental teaching of medical laboratory technology. The experiment teaching was dividedthree phases (traditional classroom teaching + simulated clinical laboratory training + clinical laboratory teaching). Classroom teaching focused on the cultivation of students' basic skills test; simulation laboratory training focused on strengthening the students' operating skills and basic post ability and professional quality; hospital laboratory teaching focused on strengthening the students' ability of medical laboratory management, test comprehensive analysis and sci-entific research. The reform practice of two sessions of the students showed that the students' examination skills and comprehensive quality improved significantly, which could better adapt to the social demand for technical personnel. Key words: three-step experiment teaching mode; Medical laboratory technology; Simulated clinical laboratory; Clinical laboratory teaching

Medical laboratory

10.3760/cma.j.issn.2095-1485.2016.05.014

Cite

Citations (0)