Ziyuan Zhong

I am currently a 5th year PhD student at the Department of Computer Science at Columbia University. I am advised by Prof. Baishakhi Ray. Previously, I did my undergrads at Reed College and Columbia University.

Email / Google Scholar / Github

Research

My current research mainly focuses on testing/improving robustness of deep learning based systems like Autonomous Driving Systems (ADSs) and Image Classifiers. Previously, I have also worked on fairness of machine learning. Representative papers are highlighted.

	CAMEO: A Causal Transfer Learning Approach for Performance Optimization of Configurable Computer Systems Md Shahriar Iqbal, Ziyuan Zhong, Iftakhar Ahmad, Baishakhi Ray, Pooyan Jamshidi SoCC, 2023 paper / bibtex We propose a method that sidesteps the limitations of existing work on system performance optimization under environmental changes by identifying invariant causal predictors under environmental changes, enabling the optimization process to operate on a reduced search space, leading to faster system performance optimization.
	Language-Guided Traffic Simulation via Scene-Level Diffusion Ziyuan Zhong, Davis Rempe, Yuxiao Chen, Boris Ivanovic, Yulong Cao, Danfei Xu, Marco Pavone, Baishakhi Ray CoRL (Oral Presentation), 2023 paper / code / bibtex We leveraged Large Langauge Models (LLMs) and developed a language guided scene-level conditional diffusion model for language-controllable, realistic traffic generation.
	Guided Conditional Diffusion for Controllable Traffic Simulation Ziyuan Zhong, Davis Rempe, Danfei Xu, Yuxiao Chen, Sushant Veer, Tong Che, Baishakhi Ray, Marco Pavone ICRA, 2023 paper / code / bibtex / demo We developed a conditional diffusion model (trained on large-scale traffic flow datasets) that can generate controllable, feasible, and realistic traffic trajectories.
	Automatic Map Generation for Autonomous Driving System Testing Yun Tang, Yuan Zhou, Kairui Yang, Ziyuan Zhong, Baishakhi Ray, Yang Liu, Ping Zhang, Junbo Chen arXiv, 2022 paper / bibtex We propose a method that can automatically generate a more concise map from a given complex map to reduce redundant test cases.
	Repairing Group-Level Errors for DNNs Using Weighted Regularization Ziyuan Zhong, Yuchi Tian, Conor J Sweeney, Vicente Ordonez, Baishakhi Ray arXiv, 2022 paper / code / bibtex A series of methods based on weighted regularization for repairing group-level errors of DNNs.
	Detecting Multi-Sensor Fusion Errors in Advanced Driver-Assistance Systems Ziyuan Zhong, Zhisheng Hu, Shengjian Guo, Xinyang Zhang, Zhenyu Zhong, Baishakhi Ray ISSTA, 2022 paper / code / bibtex FusED can efficiently identify fusion errors.
	A Survey on Scenario-Based Testing for Automated Driving Systems in High-Fidelity Simulation Ziyuan Zhong, Yun Tang, Yuan Zhou, Vania de Oliveira Neves, Yang Liu, Baishakhi Ray arXiv, 2021 paper / bibtex A Survey on Scenario-Based Testing for Automated Driving Systems in High-Fidelity Simulation.
	Neural Network Guided Evolutionary Fuzzing for Finding Traffic Violations of Autonomous Vehicles Ziyuan Zhong, Gail Kaiser, Baishakhi Ray Transactions on Software Engineering (TSE), 2022 paper / code / bibtex AutoFuzz uses a grammar-based, learning-guided fuzzing technique to efficiently find violations of Autonomous Driving Systems.
	Understanding local robustness of deep neural networks under natural variations Ziyuan Zhong, Yuchi Tian, Baishakhi Ray FASE, 2021 paper / code / bibtex DeepRobust can identify the input images whose small variations may lead to erroneous DNN behaviors.
	Testing DNN image classifiers for confusion & bias errors Yuchi Tian, Ziyuan Zhong, Vicente Ordonez, Gail Kaiser, Baishakhi Ray ICSE, 2020 paper / code / bibtex We developed a testing technique, DeepInspect, to automatically detect class-based confusion and bias errors in DNN-driven image classification software.
	Metric learning for adversarial robustness Chengzhi Mao, Ziyuan Zhong, Junfeng Yang, Carl Vondrick, Baishakhi Ray Neurips, 2019 paper / code / bibtex We propose to regularize the representation space under adversarial attack with metric learning to produce more robust classifiers.
	Noise-tolerant fair classification Alexandre Lamy, Ziyuan Zhong, Aditya Krishna Menon, Nakul Verma Neurips, 2019 paper / code / bibtex We showed both theoretically and empirically that even under the very general MC learning noise model on the sensitive feature, fairness can still be preserved by scaling the input unfairness tolerance parameter.
	Phasepack: A phase retrieval library Rohan Chandra, Ziyuan Zhong, Justin Hontz, Val McCulloch, Christoph Studer, Tom Goldstein Asilomar Conference on Signals, Systems, and Computers, 2017 paper / code / bibtex PhasePack is a collection of sub-routines for solving classical phase retrieval problems. PhasePack contains implementations of both classical and contemporary phase retrieval routines.

Misc

A Tutorial on Fairness in Machine Learning

Service

Reviewer: Neurips2020-2023, ICLR2021-2023, ICML2021-2024, ICRA2023, TOSEM

Lead Teaching Assistant: COMS 4115 Programming Language & Translators(23 Fall)

Teaching Assistant: COMS 4115 Programming Language & Translators(23 Spring), COMS 4771 Machine Learning(18 Summer, 19 Spring), ELEN 4903 Machine Learning(Edx)(18 Spring)

Design and source code from Jon Barron's website