Hao Cheng | The Hong Kong University of Science and Technology, Guangzhou

About Me

Hello! I am a Ph.D. student at at the Hong Kong University of Science and Technology (HKUST), Guangzhou. I am interested in trustworthy, efficient AI. My current research primarily focuses on ensuring the trustworthiness, robustness, and efficiency of various AI models, which is including:

Generative AI: Diffusion Models, Large Language Models, Multimodal Large Language Models, Large Audio Language Models
Embodied AI: Vision Language Action Models

Education:

the Hong Kong University of Science and Technology, Cross-campus Study Ph.D., 08/2025 - 02/2026, HongKong, China SAR
the Hong Kong University of Science and Technology, Guangzhou, Ph.D., 09/2022 - 05/2026, Guangzhou, China
Northeastern University, Ph.D., 09/2020 - 09/2022, Boston, USA
Xi’an Jiaotong University, Master, 09/2017 - 07/2020, Xi’an, China
Hebei University of Technology, B.E., 09/2013 - 06/2017, Tianjing, China

Working Experimence:

Beijing Innovation Center of Humanoid Robotics Co., Ltd., Research Intern., 07/2025 - now, Beijing, China
University of Oxford, Research Intern., 11/2023 - now, Online
SAMSUNG AI Research Center(SARC), Xi’an, Research Intern, 04/2021 - 09/2021, Xi’an, China
Alibaba DAMO Lab-Itelligent Computing Lab, Research Intern., 08/2020 - 03/2021, Shanghai, China
Northeastern University, Research Intern, 01/2019 - 06/2019, Boston, USA
National University of Singapore, Research Intern, 06/2018 - 08/2018, Singapore

News

- [07/2025] One paper is accepted to ACM MM 2025.
- [06/2025] One paper is accepted to ICCV 2025.
- [06/2025] One paper is accepted to IJCAI 2025 Workshop on Deepfake Detection, Localization, and Interpretability and is nominated as best paper.
- [06/2025] I will serve as a reviewer of NeurIPS 2025, WACV 2025.
- [06/2025] One paper is accepted to IROS 2025.
- [04/2025] I will serve as a reviewer of ACM MM 2025, IROS 2025.
- [03/2025] One paper is accepted to ICASSP 2025; One paper is accepted to ICLR 2025 Workshop on Generative Models for Robot Learning.
- [01/2025] One paper is accepted to CVPR 2025.
- [11/2024] I will serve as a reviewer of CVPR 2025, ICRA 2025.
- [09/2024] One paper is accepted to NeurIPS 2024; One paper is accepted to ACM MobiCom 2024.
- [08/2024] I will serve as a reviewer of BMVC 2024, WACV 2024.
- [07/2024] Two papers are accepted to ECCV 2024.
- [06/2024] One paper is accepted to IROS 2024.
- [05/2024] One paper is accepted to ACL 2024.
- [01/2024] One paper is accepted to CVPR 2024.
- [12/2023] One paper is accepted to WACV 2024; One paper is accepted to IEEE Transactions on Artificial Intelligence
- [08/2023] One paper is accepted to BMVC 2023.
- [05/2023] One paper is accepted to IJCAI 2023.

Publications

Preprint

Jailbreak-AudioBench: In-Depth Evaluation and Analysis of Jailbreak Threats for Large Audio Language Models

Hao Cheng*, Erjia Xiao*, Jing Shao*, Yichi Wang, Le Yang, Chao Shen, Philip Torr, Jindong Gu†, Renjing Xu†

Preprint, 2025.

PDF Code

ACM MM

Transfer Attack for Bad and Good: Explain and Boost Adversarial Transferability across Multimodal Large Language Models

Hao Cheng*, Erjia Xiao*, Jiayan Yang, Jinhao Duan, Yichi Wang, Jiahang Cao, Qiang Zhang, Le Yang, Kaidi Xu, Jindong Gu†, Renjing Xu†

ACM International Conference on Multimedia (ACM MM), 2025.

PDF

ICCV

Truthprint: Mitigating lvlm object hallucination via latent truthful-guided pre-intervention

Jinhao Duan*, Fei Kong*, Hao Cheng, James Diffenderfer, Bhavya Kailkhura, Lichao Sun, Xiaofeng Zhu, Xiaoshuang Shi, Kaidi Xu

International Conference on Computer Vision (ICCV), 2025.

PDF

CVPR

Not Just Text: Uncovering Vision Modality Typographic Threats in Image Generation Models

Hao Cheng*, Erjia Xiao*, Jiayan Yang, Jiahang Cao, Qiang Zhang, Jize Zhang, Kaidi Xu, Jindong Gu†, Renjing Xu†

Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR), 2025.

PDF

ECCV

Unveiling typographic deceptions: Insights of the typographic vulnerability in large vision-language models

Hao Cheng*, Erjia Xiao*, Jindong Gu, Le Yang, Jinhao Duan, Jize Zhang, Jiahang Cao, Kaidi Xu, Renjing Xu

European Conference on Computer Vision (ECCV), 2024.

PDF

IROS

Gaining the Sparse Rewards by Exploring Lottery Tickets in Spiking Neural Network

Hao Cheng, Jiahang Cao, Erjia Xiao, Mengshu Sun, Renjing Xu

[Pitch Oral] IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024.

PDF

ACL

Shifting attention to relevance: Towards the predictive uncertainty quantification of free-form large language models

Jinhao Duan, Hao Cheng, Shiqi Wang, Alex Zavalny, Chenan Wang, Renjing Xu, Bhavya Kailkhura, Kaidi Xu

Annual Meeting of the Association for Computational Linguistics (ACL), 2024.

PDF

CVPR

ACT-Diffusion: Efficient Adversarial Consistency Training for One-step Diffusion Models

Fei Kong, Jinhao Duan, Lichao Sun, Hao Cheng, Renjing Xu, Hengtao Shen, Xiaofeng Zhu, Xiaoshuang Shi, Kaidi Xu

Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024.

PDF

BMVC

RBFormer: Improve Adversarial Robustness of Transformer by Robust Bias

Hao Cheng, Jinhao Duan, Hui Li, Jiahang Cao, Ping Wang, Lyutianyang Zhang, Jize Zhang, Kaidi Xu, Renjing Xu

British Machine Vision Conference (BMVC), 2023.

PDF

IJCAI

Improve Video Representation with Temporal Adversarial Augmentation

Jinhao Duan, Quanfu Fan, Hao Cheng, Xiaoshuang Shi, Kaidi Xu

The International Joint Conferences on Artificial Intelligence (IJCAI), 2023.

PDF

ICCV

Adversarial robustness vs. model compression, or both?

Shaokai Ye*, Kaidi Xu*, Sijia Liu, Hao Cheng, Jan-Henrik Lambrechts, Huan Zhang, Aojun Zhou, Kaisheng Ma, Yanzhi Wang, Xue Lin

International Conference on Computer Vision (ICCV), 2019.

PDF

Journal

Early Termination of Intra Mode Decision Based on Most Probable Mode and SATD for H. 264/AVC encoding

Ping Wang, Hao Cheng

Journal of Signal Processing Systems, 2019

PDF

Projects

IJCAI Workshop

Exploring Typographic Visual Prompts Injection Threats in Cross-Modality Generation Models

Hao Cheng*, Erjia Xiao*, Yichi Wang, Linfeng Zhang, Kaidi Xu, Mengshu Sun, Xiaoshuai Hao†, Jindong Gu†, Renjing Xu†

[Best Paper Award] The International Joint Conferences on Artificial Intelligence Workshop on Deepfake Detection, Localization, and Interpretability (IJCAI Workshop), 2025.

PDF

WACV Workshop

More or Less (MoL): Defending against Multiple Perturbation Attacks on Deep Neural Networks through Model Ensemble and Compression

Hao Cheng, Kaidi Xu, Zhengang Li, Pu Zhao, Chenan Wang, Xue Lin, Bhavya Kailkhura, Ryan Goldhahn

IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW), 2022.

PDF

ICLR Workshop

Mixture of robust experts (more): A flexible defense against multiple perturbations

Hao Cheng, Kaidi Xu, Chenan Wang, Xue Lin, Bhavya Kailkhura, Ryan Goldhahn

ICLR Robust and reliable machine learning in the real world Workshop (ICLR Workshop), 2021.

PDF

KDD Workshop

Defending against backdoor attack on deep neural networks

Hao Cheng, Kaidi Xu, Sijia Liu, Pin-Yu Chen, Pu Zhao, Xue Lin

KDD Workshop on Adversarial Learning Methods for Machine Learning and Data Mining (KDD AdvML), 2019.

PDF

Professional Services

Conference Reviewer

Journal Reviewer

TNNLS, TAI, TCAD

Honors and Awards

Computer Engineering Fellowship owner at Northeastern University - 2020

Graduate with honor at Xi’an Jiaotong Universtiy - 2020

Student with honor, First Prize Scholarship at Xi’an Jiaotong University (2017 & 2018 & 2019)

Excellent graduation project at Hebei University of Technology - 2017

Student with honor, First Prize Scholarship at Hebei University of Technology (2014 & 2015 & 2016)