Hanbin Hong

Hanbin Hong

Ph.D. student

University of Connecticut

Biography

Hanbin Hong is a 4th-year Ph.D. candidate in Computer Science and Engineering at University of Connecticut, advised by Prof. Yuan Hong. His research interests lie in security and privacy issues in general machine learning models and generative models, especially in the development of innovative, universal, and efficient adversarial attacks and defense for ML models. He is currently working on developing powerful attack and defense methods for generative vision models.

He is now open to the job opportunities for 2025!

Download my resumé.

Interests
  • Machine Learning Security and Privacy
  • Generative Diffusion Models
Education

  • Ph.D. in CSE, 2022 - 2025 (expected)

    University of Connecticut

  • Ph.D. in CS, 2021 - 2022

    Illinois Institute of Technology

  • BSc in Hornor Science (Physics), 2014 - 2018

    Xi'an Jiaotong University

News

  • [07/12/2024] One first-author paper on certifiable black-box attack is accepted by ACM CCS 2024!
  • [06/12/2024] I am starting my summer internship as a research scientist on LLM-security at ByteDance!
  • [05/12/2024] One co-author paper on Backdoor attack on LLM-based code generation model is accepted by USENIX 2024!
  • [12/05/2023] I was invited as the reviewer for CVPR 2024 and ICML 2024!
  • [08/27/2023] I was invited as the reviewer for ICLR 2024.
  • [07/20/2023] One co-author paper on Certified Defense against NLP Attacks was accepted by Oakland S&P 2024! See you in San Francisco!
  • [06/13/2023] I was glad to receive the General Electric Graduate Fellowship for Excellence!
  • [05/10/2023] I was glad to receive the Predoctoral Fellowship with Excellent Research Award!
More News

[02/05/2023] I was invited as the reviewer for ICCV 2023!
[01/28/2023] I was invited as the reviewer for ICML 2023!
[01/26/2023] One paper submitted to ICML 2023!
[01/19/2023] One paper submitted to ACM CCS 2023!
[12/25/2023] I was invited as the reviewer for TDSC 2022!
[11/20/2022] I was invited as the subreviewer for AAMAS 2023! [11/01/2022] I was invited as the CVPR 2023 reviewer! [10/28/2022] I was awarded the Synchrony Financial Cybersecurity Graduate Fellowship for the Spring 2023 semester! [10/12/2022] I was awarded the student travel grant to attend CCS'22!
[07/31/2022] I was invited to serve as a PC member for AAAI-23!
[07/16/2022] A paper on Local Differential Privacy for Location-based services is accepted by ACM CCS 2022!
[07/03/2022] A paper on Universally Approximated Certified Robustness via Randomized Smoothing is accepted by ECCV 2022!
[05/31/2022] I am transfering to University of Connecticut with Prof. Yuan Hong!
[01/01/2020] A paper on Preserving Privacy Attributes in Images is accepted by ICPR 2020!

Featured Publications

Hanbin Hong, Yuan Hong
July, 2022 In arXiv 2022
Certified Adversarial Robustness via Anisotropic Randomized Smoothing
Certified Adversarial Robustness via Anisotropic Randomized Smoothing

We propose the first anisotropic randomized smoothing method which ensures provable robustness guarantee based on pixel-wise noise distributions.
PDF Cite
Hanbin Hong, Binghui Wang, Yuan Hong
July, 2022 In ECCV2022
UniCR: Universally Approximated Certified Robustness via Randomized Smoothing
UniCR: Universally Approximated Certified Robustness via Randomized Smoothing

We propose the first randomized smoothing-based universal certified robustness approximation framework against any ℓp perturbations with any continuous noise PDF.
PDF Cite code

Publications

Quickly discover relevant content by filtering publications.
Xinyu Zhang, Hanbin Hong, Yuan Hong, Peng Huang, Binghui Wang, Zhongjie Ba, Kui Ren (2023). Text-CRS: A Generalized Certified Robustness Framework against Textual Adversarial Attacks. In IEEE S&P 2024.

Cite

Hanbin Hong, Yuan Hong (2023). Certifiable Black-Box Attack: Ensuring Provably Successful Attack for Adversarial Examples. In Arxiv.

Cite

Han Wang, Hanbin Hong, Li Xiong, Zhan Qin, Yuan Hong (2022). PrivLBS: Local Differential Privacy for Location-Based Services with Staircase Randomized Response. In ACM CCS 2022.

PDF Cite

Hanbin Hong, Yuan Hong (2022). Certified Adversarial Robustness via Anisotropic Randomized Smoothing. In arXiv 2022.

PDF Cite

Hanbin Hong, Binghui Wang, Yuan Hong (2022). UniCR: Universally Approximated Certified Robustness via Randomized Smoothing. In ECCV2022.

PDF Cite code

Hanbin Hong, Yuan Hong, Yu Kong (2022). An Eye for an Eye: Defending against Gradient-based Attacks with Gradients. In arXiv 2022.

PDF Cite

Junwen Chen, Haiting Hao, Hanbin Hong, Yu Kong (2020). Rit-18: A novel dataset for compositional group activity understanding. In CVPR Workshops 2020.

PDF Cite code

Hanbin Hong, Wentao Bao, Yuan Hong, Yu Kong (2020). Privacy Attributes-aware Message Passing Neural Network for Visual Privacy Attributes Classification. In ICPR2020.

PDF Cite code

Conference Review

  • International Conference on Machine Learning (ICML) [2024]
  • Computer Vision and Pattern Recognition Conference (CVPR) [2024]
  • The Twelfth International Conference on Learning Representations (ICLR)[2024]
  • ACM Conference on Computer and Communications Security (CCS) [2023 external]
  • USENIX Security Symposium (USENIX) [2023b external]
more

- International Conference on Computer Vision (ICCV) [2023]
- International Conference on Machine Learning (ICML) [2023]
- International World Wide Web Conference 2023 (Web) [2023]
- International Conference on Autonomous Agents and Multiagent Systems (AAMAS) [2023 external]
- Computer Vision and Pattern Recognition Conference (CVPR) [2023]
- USENIX Security Symposium (USENIX) [2023 external]
- Association for the Advancement of Artificial Intelligence (AAAI) [2023 PC]
- IEEE Transactions on Dependable and Secure Computing (TDSC) [2022]
- ACM Conference on Computer and Communications Security (CCS) [2022 external]
- Conference on Neural Information Processing Systems (NeurIPS) [2022]
- International Conference on Machine Learning (ICML) [2022]
- European Conference on Computer Vision (ECCV) [2022]
- USENIX Security Symposium (USENIX) [2022 external]
- International Symposium on Research in Attacks, Intrusions and Defenses (RAID) [2022 external]
- Special Interest Group on Knowledge Discovery and Data Mining (KDD) [2022 external]
- Association for the Advancement of Artificial Intelligence (AAAI) [2022 PC]

Contact