Xiaoqian Liu

Xiaoqian Liu

Auto-generated profile, based on publicly available information

Xiaoqian Liu

Northeastern University, Shenyang, Liaoning, China

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 January 2021
Verified
Conference Paper
Open access
PDF file
XCXu Chen
Xu Chen
XLXiaoqian Liu
Xiaoqian Liu
XLXiaowen Liu
Xiaowen Liu
TWTiger Wang
Tiger Wang
CHCanan Huang
Canan Huang
JZJingbo Zhu
Jingbo Zhu
Chen Xu, Xiaoqian Liu, Xiaowen Liu, Tiger Wang, Canan Huang, Tong Xiao, Jingbo Zhu. Proceedings of the 18th International Conference on Spoken Language Translation (IWSLT 2021). 2021.

21 April 2025 in IEEE Transactions on Audio, Speech and Language Processing
Verified
Journal Article
Closed access
CXChen Xu
Chen Xu
XLXiaoqian Liu
Xiaoqian Liu
YZYuhao Zhang
Yuhao Zhang
AMAnxiang Ma
Anxiang Ma
TXTong Xiao
Tong Xiao
WYWu Yang
Wu Yang
Speech-to-text (S2T) generation tasks often struggle to achieve satisfactory convergence without relying on auxiliary data or models. We identify the core issue as the modeling granularity, where the fine-grained and lengthy characteristics of audio features pose obstacles in effectively allocating attention weights, particularly during encoder self-attention learning. In this paper, we investigate two well-established methods, Conformer and information aggregation, to mitigate the learning burden of the encoder from the aspects of intra-layer and inter-layer encoding. Conformer directly enhances modeling capability through architecture improvement, while aggregation generates coarser-grained representations, thus shaping text-like structures to fundamentally simplify attention learning. Extensive results demonstrate superior convergence and notable improvement on two representative S2T generation tasks, speech recognition and translation. In particular, we achieve a notable average BLEU score of 26.9 on MuST-C speech translation datasets without auxiliary resources and approaches. Furthermore, our finding suggests that the grown model capacity and sufficient training can effectively facilitate the granularity challenge.

5 November 2023
Verified
Conference Paper
Closed access
Xiaopan LyuXiaopan Lyu
Xiaopan Lyu
XLXiaoqian Liu
Xiaoqian Liu
RPRao Peng
Rao Peng
CYChuanzhi Yang
Chuanzhi Yang
XYXinguo Yu
Xinguo Yu
Word annotation has been adopted as a crucial preprocessing step in developing algorithms for solving arithmetic word problems (AWPs). This paper proposes a mathematical concepts enhanced annotation (MONA) model for Chinese arithmetic word problems (AWPs), consisting of two modules: mathematical Chinese word segmentation and mathematical part-of-speech tagging. The MONA model utilizes mathematical concepts to enhance the training tasks of both modules, improving their capabilities to ensure high-quality AWPs annotations. The mathematical Chinese word segmentation module is trained on an AWP corpus segmented based on mathematical concepts, resulting in higher accuracy in identifying mathematical concepts compared to general segmentation models. The mathematical part-of-speech tagging module redefines part-of-speech categories into seven classes according to mathematical concepts and trains the MONA model on the redefined part-of-speech tags, allowing more suitable part-of-speech tagging for acquiring math relations when solving AWPs. Experimental results show the proposed MONA model outperforms baseline models on benchmark AWP datasets.

1 November 2016 in China Journal of Chinese Materia Medica
Verified
Journal Article
Closed access
LYLi Yao
Li Yao
HYHong Yi
Hong Yi
HGHui-Min Gao
Hui-Min Gao
HMHui-Wen Ma
Hui-Wen Ma
JZJi-Yuan Zhang
Ji-Yuan Zhang
ZWZhimin Wang
Zhimin Wang
To compare the contents of alkaloids in theroots of cultivated and the wild Sophora flavsecens, 22 cultivated and 17 wild samples were collected. HPLC method was employed to simultaneously determine the contents of six alkaloids (oxymatrine, oxysophocarpine, sophoridine, N-methylcytisine, matrine, and sophocarpine). Independent t-test, hierarchical clustering analysis(HCA)and principal components analysis (PCA) were applied to analyze and evaluate the cultivated and the wild S.flavsecens. With a great wide range of the inter-group, the t-test results showed that the contents difference of N-methylcytisine, matrine, and sophocarpine were statistical significance(matrineandsophocarpine P<0.05, N-methylcytisine P<0.01).However, it was not statistically significant for oxymatrine, oxysophocarpine, and sophoridine.HCA and PCA showed that there were no significant differences in the contents of alkaloids of cultivated and wild S. flavsecens. The results indicated that there were no differences in the contents of alkaloids of cultivated and wild S. flavsecens.

5 November 2023
Verified
Conference Paper
Closed access
XYXinguo Yu
Xinguo Yu
BXBenyi Xie
Benyi Xie
Xiaopan LyuXiaopan Lyu
Xiaopan Lyu
XLXiaoqian Liu
Xiaoqian Liu
This paper proposes a framework for mining innovation points of scientific papers, which divides the mining of innovation points into two main sequential modules. The first module is innovation point identification, which identifies innovation points from the text of scientific papers. The second module is innovation point classification, which classifies the identified innovation points. Under this framework, this paper proposes an innovation point mining model that adopts deep learning models to guarantee the performance of innovation point identification and classification. For the innovation points identification module, this paper proposes a text convolutional neural network (Text-CNN) to identify the innovation points from the constructed statements of a given paper. For the innovation points classification module, this paper adopts a Sci-BERT to classify those identified innovation points and assign them a corresponding label. The experimental results demonstrate that the proposed innovation points mining model can effectively identify and classify innovations in scientific papers.