Interdisciplinary Research
Understanding Emotions Better With AI: Using Sound, Images, and Text Together
My research and presentation look at how combining text, audio, and visual cues improves emotion detection compared to just analyzing text. Traditional methods miss key emotional signals from tone and expressions. By analyzing all three using AI, my approach provides a more accurate understanding of emotions, based on my analysis on data from 23,000+ YouTube videos.
2024 Summer at MIT (Medlytics Track)
Lead a new team each week to best performance or #1 group, training ML on diverse medical datasets to apply AI in solving healthcare challenges, such as Hypothyroidism & sleep analysis. Below is part of my presentation of our 1st place final Capstone project.
Momentum Transfer: Moving Flea Seeds, Mouse Traps and Robots on Comets
Mentored by a professor at the University of California, this interdisciplinary research, applies the physical principle to biology, engineering, and space exploration, with the goal of enhancing the performance of future space robots.
AI in osteosarcoma cancer detection
This project utilizes AI to improve the efficacy of Osteosarcoma detection to 91% in medical images.
Housing Price Prediction Using Data Science (Python and JMP)
The project titled "Compare Python Script and JMP JSL Script on the Housing Price Regression Models" is presented at IEOM 1st Asia Pacific China Conference, July 2021. It won 3rd place of IEOM high school competition & 4th place of Stanford Online High School STEAM competition.
NeuroSupport for Autism (Passion Project)
AI chatbot for autism therapy that interacts with special needs children, enhancing communication, cognitive empathy, and socio-emotional skills through personalized conversations and exercises. It Interact with LLMs using prompt engineering.
Journalism Bias
I am creating a website which aggregates breaking news from various sources across the media landscape, covering a wide array of potentially biased sources. By using Language Models to flag potential biases, and encouraging users to view diverse viewpoints on any given news topic, the website serves to provide diverse viewpoints, and allows users to observe the biases of their preferred news platforms in an objective fashion.