2023-now

  • Assessing Country Reputation during the COVID-19 Pandemic Period on Twitter

     Country reputation is the aggregate of stakeholder's images of a country over time. Any organization or person has a reputation. The research of reputation originated from corporate reputation research in which reputation is regarded as a strategic resource that is built on numerous corporate images and actions. Researchers expanded this concept to the national level and developed a concept of country reputation. Country reputation is also a multifaceted concept that possesses a multidimensional nature. Scholars developed various multidimensional frameworks to measure this concept. To advance the measurement of country reputation, this research project will develop a comprehensive conceptual framework that consists of multiple dimensions of this concept. This framework integrates the dimensions of country reputation in the literature, as well as the dimensions of other closely related concepts such as country image and country brand. Based on this comprehensive framework, this research project will assess China's reputation during the COVID-19 period using the data of Twitter. The media reputation of a country should be an overall evaluation of multiple aspects of media coverage including media favorability (media valence measure), media visibility (media salience measure), and recency (the measure of how recent a media message is). This research project will assess the overall evaluation of social media coverage of China using the composite measure. 

2022-now

  • NSF HSI Implementation and Evaluation Project: Develop a High Quality Academic Environment for Broadening Participation of Hispanic Students in Computing

     Increasing participation of students in Computer Science (CS) and Data Science (DS) programs is an ideal way for broadening participation of Hispanic students in computing. However, the undergraduate CS and DS programs at the University of North Texas (UNT) has only about 18%, and 8% Hispanic students, respectively, which are significantly lower than 26.5% Hispanic students at UNT. This project will implement asset-based initiatives for supporting the success and broadening participation of Hispanic students in CS and DS programs. The initiatives include: (1) Infusion of computational thinking and professional skills development into the curriculum of undergraduate CS and DS programs via innovative curriculum development and mentoring; (2) Building an effective platform for supporting students' interactions with peers and faculty; (3) Developing active learning, and personalizing teaching and learning in formal as well as informal settings; and, (4) Working with education experts to foster diverse and inclusive education culture. The overarching goal of the proposed project is to enhance the computing education quality and significantly increase the enrollments, and retention and graduation rates of undergraduate Hispanic students in computing programs.

2021-now

  • Building a knowledge-driven legal IR, automatic evidence and argument generation system for decision making

     Legal information retrieval and data mining have been prominent and ongoing research no matter in data science, computer science, or legal field. Researchers in this community are trying to develop efficient artificial intelligence technology for computational law and build smart legal systems, and there is still much room to improve. In this project, we target in three useful and essential tasks in computational law: legal information retrieval, legal evidence building, and argument generation. We seek to integrate knowledge graphs and deep learning algorithms for supporting these tasks, indeed promoting service for ordinary users, legal librarians and researchers, lawyers, and judicial communities.

2018

  • Toward Reliable Disease Treatment: Effective Precision Medicine Information Retrieval (College of Information seed funding, Team Member)

     The Precision Medicine Track organized by TREC provides test collections and evaluates the performance of PM IR techniques for finding reliable medical evidence and eligible treatments. Previous TREC evaluations have significantly improved PM IR system performance. However, the performance is still far from satisfactory. For examples, the best systems conducting medical scientific abstracts retrieval could only achieve 0.4593, 0.2987, and 0.6310 in terms of IR measures infNDCG, R-prec, and P@10; As for the clinical trials retrieval task, the best results on P@5, P@10, and P@15 are around 0.5500, 0.4429, and 0.3881. There is still much room for further improvement. Our participation in TREC PM Track this year aims to explore more effective PM IR techniques. We plan to review and design innovative strategies on document indexing, query construction (query expansion), matching, and re-ranking to achieve improved retrieval performance

  • Diversifying Citation Recommendation based on Semantic Analysis of Scholar Text (NNSFC, Team Member)

     People commonly encounter three major issues when employing a citation recommendation system: 1) top-ranked recommendations are often highly similar with each other (we name it as homogenous recommendation), making it difficult to obtain sufficient literatures with a full coverage of all potential subtopics; 2) without a proper diversification mechanism, the homogenous recommendations may easily deviate from the information needs of tail users; and 3) users' needs are contextual-dependent, highly relying on their current task-at-hand, search purposes, and etc. This cannot be simply satisfied with the homogenous recommendations.As a result, this project attempts to propose a framework for citation recommendation diversification, which includes the investigation of corresponding user needs, potential application domains, dimensions for diversification, and evaluation techniques for diversified recommendations. Built on top of the past successes on search result diversification and recommendation system diversification, we are planning to expand and extend related theories/practices into the area of citation recommendation. Moreover, we would like to push the current content recommendation technique from simple text analysis toward semantic understanding; thus, we take into account multiple semantic dimensions for recommendation such as topic analysis, citation functions, and word/term functions. Our ultimate goal is to integrate the above-mentioned factors and construct a highly semantic-based, contextual-aware citation recommendation system for satisfying diversified recommendation needs. We plan to deploy a live recommendation system on a selective domain, for the purpose of exploring potential issues, such as algorithm scalability, when constructing such system, as well as evaluating the validity of our propose framework

2017

  • Multi-Semantic Information Fusion based Citation Recommendation (NNSFC, Team Member)

     Scholars may encounter the following problems when writing scientific documents: failing to find adequate related work instantly for a given research topic, retrieving appropriate citations to support the research idea, or finding that a research problem has been well established after you spent a lot of efforts on that. This project aims at constructing a theoretical framework of citation recommendation, by defining this task thoroughly and concluding the potential applications comprehensively. The project will be solved from the perspectives of both global citations and local citations, analyzing the factors like citation function (research basis, background, etc.), terminology function (problem, method, tool, etc.) and the logical structure of academic literature (introduction, relate work, method, experiment, conclusion, etc.). As a result, a citation recommendation technique is expected by fusing the above multiple semantic information. In the meantime, the project team will build a citation recommendation prototype system on a certain research domain such as computer science.