Yuliang(Leo) Chen

Yuliang(Leo) Chen

Fist-year Data Science Master Student

Yuliang(Leo) Chen

I am a second-year Data Science Master’s student at UC San Diego. My research interests lie at the intersection of machine learning, computer vision, and natural language processing. I focus primarily on multi-sensor fusion for foundation models, language-driven representation learning, self-supervised learning, and physiological signal analysis.

You can find my CV here: Yuliang Chen’s Curriculum Vitae.

Email / Github / LinkedIn

Recent News

  • [Oct 2024] Submitted a paper to ICLR 2025 for review with Yunfei!
  • [July 2024] Started working as a graduate student researcher under the supervision of Jingjing Zou on an NIH-funded project!
  • [March 2024] Joined MOSAIC Lab as a Research Assistant, collaborating with Yunfei Luo under the advisement of Professor Tauhidur Rahman!
  • [September 2023] Began the Data Science Master’s program at UCSD!
  • [October 2022] Started working as a Supply Chain Analyst at Micro Ingredients!
  • [June 2022] Graduated from UCSB with a dual major in Mathematics and Statistics-Data Science!

Publication

Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals. Yunfei Luo, Yuliang Chen (Co-first author), Asif Salekin, Tauhidur Rahman. (submitted to ICLR 2025)

Projects

Examining Longitudinal Changes in Accelerometer-Measured Physical Activity in Preventing Cardiovascular Disease with Novel Function Data Analysis Approaches (Funded by NIH)

Summer 2024
Advisor: Jingjing Zou

Vivid Panels: Deep Neural Networks for Manga Colorization

Spring 2024
Explored fine-tuning pre-trained GAN-based models for manga colorization, highlighting the performance gains achieved by addressing distribution differences between task-specific inputs and pre-training data.

VitT: Vision-Topological Transformer for Medical Image Classification

Spring 2024
Introduced the VitT model, strengthening Vision Transformer (ViT) with lightweight task-specific encoder and fusion layer, resulting in better performance with minimal computational cost.

E-StyTR^2: Efficient Image Style Transfer with Transformers

Spring 2024
Investigated various fusion modules based on StyTr2 to effectively blend style and content, evaluating their efficiency and aesthetic quality using quantitative metrics.

MAE-Fundus: Foundation Model On Retinal Images using Masked Autoencoders

Fall 2023,Winter 2024
Developed foundation model for fundus image analysis using Masked Autoencoders (MAE) under the supervision of Professor Pengtao Xie.

Image-to-Image Retrieval with CLIP

Winter 2024
Developed an image-to-image retrieval system using CLIP, demonstrating its superior ability to capture robust and generalized image representations compared to traditional CNNs like ResNet.

Fall 2023
Investigated the challenges and strategies related to explanation multiplicity in machine learning models, highlighting the implications of multiple explanations for model decisions and developing effective methods to manage them.