Yunjuan Wang / 汪云娟

I'm a theoretical Machine Learning Ph.D. at Department of Computer Science, Johns Hopkins University. I am fortunate to be advised by Raman Arora.

My research interest focuses on Trustworthy AI (Robust Adversarial Learning), Deep Learning Theory, Transfer Learning, Theoretical ML, etc. I'm passionate about understanding the fundamental question of machine learning and deep learning, as well as using ML algorithms and data-driven techniques to solve real-world problems.


Work Experience

Machine Learning Engineer Intern

Meta (Menlo Park)

Train multi-modality foundational models for AI integrity purposes.

May 2024 - Aug 2024

Student Researcher

Google Research (New York)

Designed practical defense algorithm against adversarial attack for unsupervised domain adaptation that outperform competitive baselines and along with theoretical guarantees.

Jun 2023 - Sep 2023

Student Researcher

Google Research (Seattle)

Designed Importance Weighted Subset Selection algorithm (IWeS) that outperforms competitive baselines on large datasets and along with theoretical guarantees.

May 2022 - Aug 2022

Research Experience

Graduate Research Assistant

Johns Hopkins University

Responsible for designing defense algorithms in DARPA project on AI Robustness against deception.

Explored the foundation of robust adversarial learning.

Sep 2019 - Present

Undergraduate Research Assistant

University of Illinois at Chicago

Investigate bandit problem under Fatigue-aware Online Recommendation System.

Sep 2018 - Apr 2019

Teaching Experience

EN.601.475/675 Introduction to Machine Learning

Spring 2021, Spring 2023, Fall 2023

EN.601.779 Machine Learning: Advanced Topics

Spring 2022

Publications / Preprints

On the Stability and Generalization of Meta-Learning

Yunjuan Wang, Raman Arora

Advances in Neural Information Processing Systems (NeurIPS), 2024

Stability and Generalization of Adversarial Training for Shallow Neural Networks with Smooth Activation

Kaibo Zhang, Yunjuan Wang, Raman Arora

Advances in Neural Information Processing Systems (NeurIPS), 2024

DART: A Principled Approach to Adversarially Robust Unsupervised Domain Adaptation

Yunjuan Wang, Hussein Hazimeh, Natalia Ponomareva, Alexey Kurakin, Ibrahim Hammoud, Raman Arora

In submission

Adversarially Robust Hypothesis Transfer Learning

Yunjuan Wang, Raman Arora

International Conference on Machine Learning (ICML), 2024

Benign Overfitting in Adversarially Trained Neural Networks

Yunjuan Wang, Kaibo Zhang, Raman Arora

International Conference on Machine Learning (ICML), 2024

Leveraging Importance Weights in Subset Selection

Gui Citovsky, Giulia DeSalvo, Srikumar Ramalingam, Afshin Rostamizadeh, Yunjuan Wang (alphabetical order)

International Conference on Learning Representations (ICLR), 2023

Adversarial Robustness is at Odds with Lazy Training

Yunjuan Wang, Enayat Ullah, Poorya Mianjy, Raman Arora

Advances in Neural Information Processing Systems (NeurIPS), 2022

Robust Learning for Data Poisoning Attacks

Yunjuan Wang, Poorya Mianjy, Raman Arora

International Conference on Machine Learning (ICML), 2021

Thompson sampling for a fatigue-aware online recommendation system

Yunjuan Wang, Theja Tulabandhula

Workshop on Information Technologies and Systems (WITS), 2021

Base belief function: an efficient method of conflict management

Yunjuan Wang, Kezhen Zhang, Yong Deng

Journal of Ambient Intelligence and Humanized Computing, 2018


Education

Johns Hopkins University

PhD student
Computer Science

GPA: 4.0/4.0

Sep 2019 - Expected 2025 Spring

University of Illinois at Chicago

Visiting student
Electrical and Computer Engineering

GPA: 4.0/4.0

Sep 2018 - May 2019

Xi'an Jiaotong University

Bachelor of Science
Computer Science

GPA: 3.89/4.3

Sep 2015 - Jun 2019

Service

Conference Reviewer

International Conference on Machine Learning (ICML), 2022 - 2023
Conference on Neural Information Processing Systems (NeurIPS), 2022 - 2023
International Conference on Learning Representations (ICLR), 2023 - 2024
International Conference on Artificial Intelligence and Statistics (AISTATS), 2023

Program Committee Member

AAAI Conference on Artificial Intelligence, 2024