top of page

Access my Curriculum Vitae and resume.

Research Grants

2022-2023

Meta PhD Fellowship Finalist (110 of 3200)

UCLA Graduate Council Diversity Fellowship (full tuition and stipend coverge)

2023-2024

2023 - 2024

CCSE Research Grant

2023

UCLA ISH Research Innovation Fund Finalist

2023

UCLA Graduate Student Research Mentorship Program                                                        

2023

UCLA Canadian Studies Graduate Student Research Award 

Awards & Honors

August 2022

2022 AEJMC Annual Conference, Top Method Paper

August 2022

2022 AEJMC Annual Conference, First Place Student Paper

August 2021

2021 AEJMC Annual Conference Master's AWARD

August 2021

2021 AEJMC Annual Conference, Third Place Student Paper

2015 - 2018

USC Marshall Dean’s List

Research Experiences

August 2020 - December 2021

Communicative Mind (Co-Mind) Lab, UCLA

Principal Investigator: Dr Rick Dale, UCLA

  • Developed various multimodal architectures (RNN, CCN, autoencoder, S2S) integrating OpenPose, MFCC, and BERT embeddings to simulate brain-like information processing, introducing multiple supervised and unsupervised techniques for generating multimodal representations

  • Investigated correlations between distinct representational learning types (e.g., AI vs. behavioral vs. neurosignals, symbolic vs. numerical, unimodal vs. multimodal), enabling manipulation of representational systems for understanding human vs. AI cognition

  • Proposed several statistical approaches (logit, SVM, KNN, t-SNE) to interpolate the learning and information processing trajectory of deep learning models

August 2020 - December 2021

Artificial Intelligence and Emerging Media Lab, Boston University

Principal Investigator: Dr Lei Guo, BU

Project: Communicating Covid-19, methodology research on LDA topic modeling and news framing

  • Designed a stable scraping method that overcame data deficiency and copywrite issues encountered in the existing news archives, as well as scraping tools for team to get articles from 40+ news websites

  • Modified an LDA model for underrepresented Ugandan coverage, optimizing topic analysis in this context by adding religion, law and justice, and international aid to Africa as sub-topics to the worldwide topic list

  • Collaborated with a Ugandan coder to complete an intercoder reliability test for content analysis, coded 100 stratihied news samples and conducted content analysis of 1,000 news samples to prepare a training dataset for Ugandan news framing annotation

  • Conducted news framing research as part of developing a framing element detection tool in collaboration with the computer science and communication faculties

April 2020 - December 2021

Education Department, Hong Kong Baptist University

Principal Investigator: Dr Wai-Chi Chee, HKBU

Project:  Understanding the sociopolitical participation of ethnic minority youth in Hong Kong

Funder: Public Policy Research (PPR) Funding Scheme under Hong Kong Policy Innovation and Co-ordination Office (PICO), (ID: SR2020.A2.013)                                                       

  • Performed in-depth qualitative analysis based on 15-hour interviews and 140+ pages of transcripts to investigate the barriers and facilitators of ethnic minority (EM) youth, the roles they take up, and potential policy solutions to enhance their participation

  • Analyzed interviews using an anthropology framework of theme coding based on parent- and sub-level nodes, improving the efhiciency of analysis by employing matrix coding in NVivo

  • Led a comprehensive survey design by comparing six political engagement scales and selecting three relevant scales for measuring the political perceptions, behaviors and aspirations of EM youth

  • Handled the full quantitative analysis of the survey dataset by applying multiple regression models, composite analysis, ANOVA, and t-tests to complement our interview analysis

 

Project:  Understanding the barrier and facilitators of first-generation university students of ethnic minorities in Hong Kong (research proposal and funding application)

Proposal submitted to PPR Funding Scheme under Hong Kong PICO               

  • Drafted the 15-page research proposal and adopted a mixed method approach by utilizing policy analysis, interviews, and surveys, framing research to investigate the dual marginalization faced by these students due to their ethnic identities and low socioeconomic status in Hong Kong

  • Adopted a “success-factor” model by viewing the ethnic identities as cultural capital instead of as a barrier, advocating for more effective policy design and policy implementation at three levels of educational practices (government, institution and classroom) for EM students in Hong Kong

April 2020 - July 2020

Whitehead Communication

Data Scientist & Communication Researcher

  • Utilized Python intensively to scrape from multiple social media platforms, including Twitter, Facebook, Tripadvisor, Google Play, and 40+ Ugandan newspaper websites to cover different data dimensions;

  • Implemented six natural language processing (NLP) models, including sentiment analysis, LDA topic modeling, word2vec text categorization, K-means clustering, to perform multiple NLP analysis on time series Facebook and Twitter data and enrich our data dimensions

  • Produced network analysis and Python tutorials in both written and video versions and held webinars to teach our team using Gephi for advanced data analysis and visualization.

White paper: Topic Modeling of Covid-19 Media Coverage in Uganda

  • Developed an optimal LDA model by looping through a different set of parameters and topic numbers from 15 to 22, and used this model to conduct topic analysis for online news under the Ugandan context

  • Published a white paper of Covid-19 news coverages in Uganda with the founder Anne Whitehead

July 2019 - January 2020

National Natural Science Foundation of China (ID:71573074)

Project: The Impact of Mergers and Acquisitions (M&A) on Acquires’ Capital Structures: Empirical Research Based on Industries, Regions, and Company Natures

  • Combined and cleaned 69,374 financial transactions and 6,099 companies’ data using Python with positive outcomes and increased empirical results accuracy by 38%;

  • Adopted difference-in-difference (DID) based on literature review to test short-term M&A effects, independently designed dynamic DID models through STATA to test longitudinal and time-lag effects;

  • Developed original STATA script instruction code for preliminary results based on multiple econometric models, improved STATA execution efficiency by 60%;

  • Led a junior research assistant to export 234 analyzable results tables, categorizing into different industries, regions, and company natures to improve the robustness of results.

bottom of page