Xuanzhou Chen

This is Xuanzhou Chen. I am currently a MS student at NYU and also a 2023 Phd applicant. My research vision is to bridge the gap between theoreties and applications in deep learning, specifically interested in building efficient/robust/generalized deep learning algorithms to be applied in multiple fields such as NLP and medical imaging. Overall, my research interests lie in the intersection of :

1. Machine Learning Privacy & Security
2. Information Theory
3. Machine Learning Theory
4. Deep Learning Algorithm
5. Data Analysis

Overall, I am eagerly exploring some research problems as following:

1. How to make deep neural network models more robust against adversarial perturbations? -> R-D theory?
2. How to generalize a deep learning model (such as a transformer-based model) by investigating inner topological strcture in neural networks and high dimensional data?
3. How to make the training process more computationally efficient? How to remove “redundant” information while training DL models?
4. How to mitigate the inherent bias in pretrained models and train them fairly in downstream tasks?

Currently, I am self-learning Topological Structures in Neural Networks, Topological Data Analysis, and Multi-task Learning.

2023 New Year Aspirations: Stay curious, read more, learn more, experiement more, socialize more, and start to draft my research proposal!