Zhitong Gao's Homepage

About Me

Hi, I am a PhD student at EPFL, working on machine learning and computer vision. I received an EDIC fellowship and am advised by Prof. Amir Zamir at VILab. Currently, I am mostly interested in building generalizable multimodal multitask models, especially considering the missing properties between current models and humans. For the long term, I am excited about unraveling how human intelligence works through the lens of machine learning and am eager to see how we can boost AI to improve human life.

Previously, I completed my master's studies at ShanghaiTech University PLUS Lab, working with Prof. Xuming He. I also have an exchange semester at EPFL's CVLab, working with Dr. Mathieu Salzmann. During my master's, I studied semantic segmentation under distribution shifts and explored topics like test-time adaptation, uncertainty estimation, and domain generalization, with tools such as optimal transport, generative modeling, and Bayesian deep learning.

Happy to chat if you share similar interests or just want to connect. 😊

News

10/2024: One paper on semantic segmentation under distribution shifts was accepted by NeurIPS 2024.
09/2024: I started my PhD studies at EPFL and joined VILab. Excited for this new journey!
05/2024: I completed my master's studies and successfully defended my thesis titled "Uncertainty Modeling and Its Applications in Semantic Segmentation."
12/2023: I attended the NeurIPS 2023 in New Orleans and gave a poster presentation.
11/2023: I received the China National Scholarship (rate 0.2%)!
11/2023: One paper was accepted by ML4H 2023. Congrats to Bingnan!
10/2023: I gave both an oral and a poster presentation at the ICCV 2023 UNCV Workshop in Paris.
09/2023: One paper was accepted by NeurIPS 2023.
09/2023: I have joined the EPFL CV lab as a visiting scholar, working with Mathieu Salzmann.
08/2023: I served as a reviewer for NeurIPS 2023, reviewed 2 papers.
08/2023: I attended the IJCAI conference onsite in Macau and gave both an oral and a poster presentation.
05/2023: I attended the ICLR conference onsite in Rwanda and gave a poster presentation.
04/2023: One paper was accepted by IJCAI 2023. Congrats to Chuanyang!
01/2023: I served as a reviewer for CVPR 2023, reviewed 2 papers.
01/2023: One paper was accepted by ICLR 2023.
10/2021: I attended the MICCAI conference online and gave a poster presentation.
07/2021: I finished my undergraduate studies and received an outstanding undergraduate thesis award!
05/2021: One paper was accepted by MICCAI 2021.

Publications

ATTA: Anomaly-aware Test-Time Adaptation for Out-of-Distribution Detection in Segmentation

[PDF] [Code] [Poster] [Slides]

Zhitong Gao, Shipeng Yan, Xuming He

Conference on Neural Information Processing Systems (NeurIPS), 2023

Out-of-Distribution Detection Anomaly Segmentation Test-Time Adaptation

Gradient-Map-Guided Adaptive Domain Generalization for Cross Modality MRI Segmentation

[PDF] [Code]

Bingnan Li, Zhitong Gao, Xuming He

Machine Learning for Health (ML4H), 2023

Domain Generalization Test-Time Adaptation Medical Image Segmentation

Modeling Multimodal Aleatoric Uncertainty in Segmentation with Mixture of Stochastic Experts

[PDF] [Code] [Poster] [Video]

Zhitong Gao, Yucong Chen, Chuyu Zhang, Xuming He

International Conference on Learning Representations (ICLR), 2023

Aleatoric Uncertainty Estimation Stochastic Segmentation Generative Model

MILD: Modeling the Instance Learning Dynamics for Learning with Noisy Labels

[PDF] [Code]

Chuanyang Hu, Shipeng Yan, Zhitong Gao, Xuming He

International Joint Conference on Artificial Intelligence (IJCAI), 2023

Robust learning Learning Dynamics

Superpixel-guided Iterative Learning from Noisy Labels for Medical Image Segmentation

[PDF] [Code] [Poster] [Video]

Shuailin Li*, Zhitong Gao*, Xuming He

International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), 2021

Robust learning Medical Image Segmentation

Note: * indicates equal contribution.

Challenges and Workshops

MUAD Uncertainty Estimation for Semantic Segmentation Challenge at ICCV2023

[Challenge Website] [Report] [Slides]

Zhitong Gao, Xuming He

Top 7; Invited for oral and poster presentation at ICCV UNCV Workshop.

Out-of-distribution Detection Uncertainty Estimation Test-Time Adaptation

Quantification of Uncertainties in Biomedical Image Quantification Challenge at MICCAI2021

[Challenge Website] [Report] [Poster]

Yucong Chen, Guanqi He, Zhitong Gao, Xuming He

Invited for oral and poster presentation at MICCAI 2021 Workshop.

Uncertainty Estimation Medical Image Segmentation

Course Projects

Understanding and Exploration for "Adaptive Online Learning in Dynamic Environments"

[Report]

Zhitong Gao*, Haotian Tian*, Tianyue Zhou*

Final Project of CS245 Online Optimization and Learning, 2023 Spring, ShanghaiTech University

Online Learning Optimization Non-stationary Environment

Generalized DUQ: Generalized Deterministic Uncertainty Quantification

[Report]

Zhitong Gao*, Ziyao Zeng*

Final Project of CS282 Machine Learning, 2021 Spring, ShanghaiTech University

Uncertainty Quantification Aleatoric Uncertainty Epistemic Uncertainty

Seek Common while Shelving Differences: A New Way for dealing with Noisy Labels

[Report]

Zhitong Gao*, Ziyao Zeng*

Final Project of CS280 Deep Learning, 2020 Fall, ShanghaiTech University

Robust Learning Noisy label

About Me

News

Publications

Challenges and Workshops

Course Projects

Services