Hi, this is Cecelia!!
I’m an enthusiastic learner, while my core motivation is
**How do human adapt to tools?**
By tools, I mean everything created by human to describe or control the world, from symbolics to AI.
By adapt, I mean more than learning to use, but also how tools and human reshape each other in such process.
This has many sub-questions and also requires building up apparatus, and each leads to new sub-questions, e.g.
- What’s the border of brain decoding?
- How to push Cog-Sci to the wild world?
- How to connect Cog-Sci with more data-driven methods?
Outside the lab, I'm always enjoying literatures/mangas/films (How creative could human be?!),
discussing with friends, and creating artworks
for my favorite characters (you might’ve come across me somewhere on the Internet :D)
Education
B.E. candidate in Computer Science, University of Science and Technology of China
Research experience
I struggled a long way to find my research interests, and I am grateful to all the mentors who guided me through the process. Here are some of my research experiences:
Exploring the Task Space of Embodied Assistance
Human Computer Integration Lab · University of Chicago
Advisor: Prof. Pedro Lopes Ongoing
- Developing visual closed-loop control for parallel electrical muscle stimulation.
- Conducting behavioral and user studies with blind and low-vision participants.
Longitudinal Brain MRI Prediction via Diffeomorphic Deep Learning
National Engineering Lab for Brain-inspired Intelligence Technology
Advisor: Prof. Aiping Liu
- Designed DRIFT, a two-stage framework for predicting follow-up brain MRIs from baseline scans using a conditional 3D U-Net to generate diffeomorphic velocity fields.
- Achieved a 78–83% MSE reduction over state-of-the-art methods on 3,926 UK Biobank longitudinal pairs, with interpretable Jacobian maps and velocity fields localizing regional atrophy.
- Independently led problem formulation, model design, all experiments, and manuscript writing.
mmWave Radar-Based Passive Sensing and Communication System
Lab for Intelligent Networking and Knowledge Engineering
Advisor: Prof. Yubo Yan
- Conducted electromagnetic simulations and parameter optimization for radar sensing; designed and validated hardware prototypes.
LLM-Empowered Memory Management Optimization Framework
Operating Systems (H) · Course project
Advisor: Prof. Kai Xing · Code ↗
- Led a five-person team to build a predictive framework using locally deployed Llama 3-8B-Instruct to anticipate user behavior for system-level memory pre-allocation, achieving approximately 70% accuracy on web tasks.
VEGFR-like Surface Display on E. coli for Tumor Inhibition
iGEM · Presented at the Grand Jamboree, Paris
Advisor: Prof. Jiong Hong
- Contributed to experimental design and data analysis; led scientific communication and the team presentation.