Alex Huang

Hi! I'm Alex Huang, a first-year PhD student at the University of Michigan–Ann Arbor, advised by Prof. Emily Mower Provost. My research lies at the intersection of speech-centered machine learning and affective computing, with a focus on developing data-efficient methods for building emotion-aware AI systems.

Before coming to Ann Arbor, I earned my Bachelor’s degree in Computer Science from UW–Madison, where I worked with Prof. Yin Li and Prof. Timothy Rogers.

Email / Scholar / LinkedIn / Github

Publications

	A Data-driven Approach to Facial Expression Recognition Wei-Chun Huang Senior Thesis Using Transformer based VaDE model to cluster facial expressions extracted using a 3D face reconstruction model (EMOCA).
	Uniting theory and data: The promise and challenge of creating an honest model of facial expression Sophie Wohltjen, Y. Ivette Colon, Zihao Zhu, Karina Miller, Wei-Chun Huang, Bilge Mutlu, Yin Li, Paula Niedenthal Cognition and Emotion paper Discuss the resources that are available to help researchers build a more ecologically valid model of facial expressions.
	Identity fusion with proximal but not distal social groups rises after mastering a coordinated marching routine Michelle Marji, Joshua Conrad Jackson, Karina J Miller, Siddharth Suresh, Wei-Chun Huang, Alexis Y Liu, Christian G Andresen, Corey Pompey, Jamin Halberstadt, Paula M Niedenthal Royal Society Open Science paper Track and measure synchrony and spatial configurations of large groups using Faster R-CNN.
	Categories vs Semantic Features: What shape the similarities people discern in photographs of objects? Siddharth Suresh, Wei-Chun Huang, Kushin Mukherjee, Timothy Rogers ICLR Workshop, 2024 paper Discovered that model trained to produce category labels and model trained to generate semantic features learned very different representational geometries throughout the network.
	Conceptual structure coheres in human cognition but not in Large Language Models Siddharth Suresh, Kushin Mukherjee, Xizheng Yu, Wei-Chun Huang, Lisa Padua, Timothy Rogers EMNLP, 2023 paper Compared how Large Language Models such as FLAN and GPT structure concepts relative to human cognition.

Other Projects

Below are some of the non-research projects I’ve worked on.

	3D Facial Feature Extraction repo Extract 3D Facial Features from videos and images using SOTA 3D face reconstruction model - EMOCA .
	Video Feature Extraction repo This directory contains the code to extract features from video datasets using mainstream vision models such as Slowfast, i3d, c3d, CLIP, etc.

This website is built using a template created by Jon Barron. You can find the source code: here .