Joie Yeahuay Wu

I am a PhD student at University of Massachussetts Amherst, where I work on representation learning from multimodal data. My PhD advisor is Ina Fiterau.

I have a B.S in Computer Science and Mathematics from Temple University. During undergrad, I was mentored by Vasily Dolgushev, Bo Ji, and Anne Ngu.

github  /  linkedin

profile photo

Research

I'm broadly interested in disentangled representation learning, and multimodal deep learning. I am interested in advances in these fields that result in improved model performance on downstream tasks such as time-series forecasting and diagnosis of neurodegenerative diseases such as Alzheimer's.

PontTuset FLARe: Forecasting by Learning Anticipated Representations
Surya Teja Devarakonda*, Joie Yeahuay Wu*, Yi Ren Fung, Madalina Fiterau
MHLC (Machine Learning for Healthcare) , 2019
project page

A novel time series forecasting approach using LSTMs which predicts representations across the forecasting horizon.

PontTuset Alzheimer's Disease Brain MRI Classification: Challenges and Insights
Yi Ren Fung*, Ziqiang Guan*, Ritesh Kumar, Joie Yeahuay Wu, Madalina Fiterau
IJCAI ARIAL Workshop, 2019
project page

Comprehensive study of Alzheimer's disease diagnosis using CNNs. Discovered data leakage issue from previous papers.