Lei Xia's Homepage

Lei Xia is currently pursuing his Ph.D. in Computer Science at Hong Kong Baptist University (HKBU), under the joint supervision of Dr. Yupeng Li (HKBU), Dr. Hong-Ning Dai (HKBU), and Prof. Francis C.M. Lau (The University of Hong Kong). He is a member of HKBU’s Research Clusters in Data Analytics and Artificial Intelligence. Lei obtained his MSc in Computer Science with distinction from The University of Hong Kong in 2024, where he was supervised by Dr. Qi Liu. Previously, he was an intern at Tencent AI Lab from July 2022 to August 2023.

His research interests include artificial intelligence, data analytics, and their applications in real-world scenarios. To be specific, he is currently interested in Large Language Models , Social Computing, and Misinformation.

Lei Xia is open to any collaboration opportunities. If you’re interested in working with him, please feel free to contact him via email.

news

Sep 26, 2024	One paper accepted by NeurIPS 2024 (D&B Track): MEQA: A Benchmark for Multi-hop Event-centric Question Answering with Explanations.
Sep 20, 2024	One paper accepted by EMNLP 2024 (Findings): Document-level Causal Relation Extraction with Knowledge-guided Binary Question Answering.
Aug 13, 2024	One paper published by FinNLP-AgentScem@IJCAI-2024: Using Pre-trained Language Model for Accurate ESG Prediction.

selected publications

FinNLP

Using Pre-trained Language Model for Accurate ESG Prediction

Lei Xia, Mingming Yang, and Qi Liu

In Proceedings of the Eighth Financial Technology and Natural Language Processing and the 1st Agent AI for Scenario Planning, 2024

Bib

@inproceedings{xia2024using,
  title = {Using Pre-trained Language Model for Accurate ESG Prediction},
  author = {Xia, Lei and Yang, Mingming and Liu, Qi},
  booktitle = {Proceedings of the Eighth Financial Technology and Natural Language Processing and the 1st Agent AI for Scenario Planning},
  pages = {1--22},
  year = {2024},
}

VINCI
BiverWordle: Visualizing Stock Market Sentiment with Financial Text Data and Trends

Lei Xia, Yi-Ping Gao, Le Lin, and 2 more authors

In Proceedings of the 16th International Symposium on Visual Information Communication and Interaction, Guangzhou, China, 2023

Abs Bib

While financial forums are increasingly significant in financial analysis, current visualization tools do not properly exploit their text data. To address this, we present BiverWordle, a novel tool that reveals the relationship between market sentiment and firm trends. BiverWordle integrates candlestick chart, ThemeRiver, and Wordle with text classification and sentiment analysis techniques to decode market dynamics from textual sources, such as shareholder opinions and firm announcements. With the application of a Voting model to the manually labeled data, we achieved an accuracy of approximately 64%. BiverWordle facilitates the extraction of shareholder insights from sparse comments and provides a visual method for historical stock trend analysis, which we validated with three distinct stock trends. Resources are accessible at https://github.com/Brian-Lei-XIA/BiverWordle.
@inproceedings{10.1145/3615522.3615541, author = {Xia, Lei and Gao, Yi-Ping and Lin, Le and Chen, Yu-Xi and Zhang, Kang}, title = {BiverWordle: Visualizing Stock Market Sentiment with Financial Text Data and Trends}, year = {2023}, isbn = {9798400707513}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3615522.3615541}, doi = {10.1145/3615522.3615541}, booktitle = {Proceedings of the 16th International Symposium on Visual Information Communication and Interaction}, articleno = {18}, numpages = {5}, keywords = {Visualization, Sentiment Analysis, Financial Text Processing}, location = {Guangzhou, China}, series = {VINCI '23}, }

news

selected publications

visitor map