Wanting Mao

毛婉婷

Logo

Senior at UC San Diego majoring in
Data Science and Mathematics(Applied).

Email: wamao@ucsd.edu

View My CV

View My LinkedIn Profile

View My GitHub Profile

About Me

I’m a fourth-year undergrad at UC San Diego pursuing B.S. degrees in Data Science and Mathematics(Applied). I’m curious about Machine Learning models and their applications in inverse problems, vision, science, and many other fields, but I’m also willing and hoping to explore many other fields in CS, DS and Mathematics. I am looking forward to gaining more about these areas from my undergrad education and beyond. My ultimate goal is to create a safer, happier, and more equitable world with all my knowledge. Borrowing the slogan of Bayer, my motto is

Science for a Better Life

Research Experience

  1. Implementation of Distribution Testing Algorithms with Professor Daniel M. Kane
    • September 2021 - Present
    • Volunteer Undergraduate Researcher at Early Research Scholars Program (ERSP)
    • Collaborating with Aoxi Li, Juyoung Park, and John Wu, I implement recent algorithms in distribution testings, test the algorithms’ efficiency, and determine whether the theoretical expectations match the experimental outcome. Repo
    • Participating in the writing of the paper “Experimental Evaluations of Distribution Testing Algorithm” and prepared for submission.
  2. From Theory to Pixels: Diffusion Models for Image and Data Generation with Professor Alex Cloninger and Professor Rayan Saab
    • September 2023 - March 2024
    • Data Science major capstone project
    • Explored limitations of diffusion model in reconstrusting the exact distributions. Also, investigated on it’s inability to reconstruct multiple distributions. Those findings provide understanding into the theoretical reasons behind generating the out-of-distribution data points.
    • In winter quarter, with my awesome team members(Jiatu Li], Zhengyun Nie, and Jessica Song), I am going to leverage the numerical accuracy of diffusion models. We noticed that diffusion model is not good at generating images with desired number of objects. Thus, we want to train diffusion model specifically on numerical accuracy and offer an interactive input methods for users.
  3. High-fidelity Customized Scene/Character Video Generation
    • Januaray 2024 - Present
    • Tencent Games Engine Technology Salon, Rocket Plan
    • Current generative models are widely used in image and video generation, however, the results usually cannot achieve high-fidelity as expected. We want to study algorithms that generate high-fidelity video with customized characters or scenes from the perspective of image dynamics.
  4. The Impact of Security-Related Chatbots on Users’ Mental Model with Professor Imani N. S. Munyaka
    • September 2022 - December 2022 and March 2023 - June 2023
    • Admitted as part of the ERSP summer program.
    • Background: Although recommendation chatbots are commonly used, we want to assess the influence of a security-focused chatbot on user mental frameworks and the overall user experience.
    • Explored SAUR Conversational Recommender System, Saleforce Converse and Rasa framework for conversational chatbot.
    • Developed Recommendation AI Chatbot for security softwares, including the design of icon and UI. Slides, Talk
  5. CSNext: Graduate School Prep Workshop at UW Reality Lab with Shirley Xue and Antonio Glenn
    • March 2023 - May 2023
    • Selected to be part of the UW CSnext workshop.
    • Processed ECG signal data collected by smart earrings.
    • Employed ML techniques(Adaboost decision tree, 1D-CNN etc.) to identify abnormal ECG signals with scikit-learn and Tensorflow keras.Slides
  6. Tweeter Data Analysis with Professor Imani N. S. Munyaka
    • June 2022 - September 2022
    • Early Research Scholars Program (ERSP) Summer Program
    • Twitter analysis including Data Gathering, Data Preprocessing, Keyword Trend Analysis, Sentiment Analysis, and Topic Modeling in both R and Python.
      • Individually analyzed 2300+ tweets mentioning “spam call”-related keywords and tags.
      • Assisted team members(Akshay Mehta, Claire Lee and Xinyi(Cindy) Wang) analyze a total of 10000+ tweets.
      • Mentored high school students on the project.

Projects

  1. Dress Recommendation Chatbot
    • In collaboration with Xinyi(Cindy) Wang, we created a simple chatbot that recommends dresses for users.
  2. Auto login tool for amazon.com
    • A script that helps users log into amazon.com automatically.
  3. Rotten Tomato Movie Score Visualization
    • This is the final project of the DSC106 @UCSD instructed by Professor Soohyun Nam Liao during Spring 2023. Git repo
    • With the help of JavaSript D3, HTML, and CSS, this project explores what elements make a good movies.
    • This visualization consists three plots with interaction, including scatter plot, geometric plot, and network plot.
  4. Analysis of US House Representatives Stock Trade
    • This is a project assignment from DSC 80 @UCSD instructed by Professor Justin Eldridge during Fall 2022.
    • With Zhengyun Nie(znie@ucsd.edu), we analyzed the stock trades of the US House of Representatives.
    • This project is consisted of two parts:
      • Data Propocessing & Analysis: We cleaned the data, explored the data through univariate and bivariate anlysis as well as aggregation, assessed the missingness of one of the columns, and conducted a hypothesis test on whether trades are evenly distributed on each weekday.
      • ML classifier: Using the clean data we obtained from the previous analysis, we built up a decision tree classifier model which predicts whether a transaction if purchare or sale given 12 features with ≈ 68 percent accurracy. Moreover, we analzed the fairness of our model with respect to the titles of different representatives.
  5. Weather Bot
    • A practice of Rasa following a tutorial.

Awards