Jun Yan

Jun Yan

Posts by Collection

portfolio

publications

Segment-anything models achieve zero-shot robustness in autonomous driving

Published in IEEE IAVVC, 2024

Best paper of IEEE IAVVC. An early work on adversarial robustness of foundation models

Recommended citation: Yan, Jun, Pengyu Wang, Danni Wang, Weiquan Huang, Daniel Watzenig, and Huilin Yin. "Segment-anything models achieve zero-shot robustness in autonomous driving." In 2024 IEEE International Automated Vehicle Validation Conference (IAVVC), pp. 1-8. IEEE, 2024.
Download Paper

Sparse Query Dense: Enhancing 3D Object Detection with Pseudo Points

Published in ACM MM, 2024

An oral and wonderful paper for safe autonomous driving.

Recommended citation: Mo, Yujian, Yan Wu, Junqiao Zhao, Zhenjie Hou, Weiquan Huang, Yinghao Hu, Jijun Wang, and Jun Yan. "Sparse Query Dense: Enhancing 3D Object Detection with Pseudo points." In Proceedings of the 32nd ACM International Conference on Multimedia, pp. 409-418. 2024.
Download Paper

Spectral Scaling-Based Augmentation for Corruption-Robust Image Classification

Published in IEEE Signal Processing Letters, 2025

Image classifiers often degrade in performance when test images differ significantly from the training distribution due to real-world image corruptions. Frequency-based augmentations can be used to address this issue, but existing methods excel against corruptions caused by noise and blur while struggling with those caused by contrast and fog. To tackle these challenges, we propose a novel image augmentation method grounded in a new perspective of relative spectral differences.

Recommended citation: Zhang, Zhuang, Lijun Zhang, Dejian Meng, Wei Tian, and Jun Yan. "Spectral Scaling-Based Augmentation for Corruption-Robust Image Classification." IEEE Signal Processing Letters (2025).
Download Paper | Download Slides

Exploring Causal Information Bottleneck for Adversarial Defense

Published in IEEE Transactions on Information Forensics and Security, 2025

This paper addresses the significant issue in IB by incorporating causal inference into the IB-based defense framework

Recommended citation: Yan, Jun, Huan Hua, Weiquan Huang, Xi Fang, Wancheng Ge, Jiancheng Yang, and Yongwei Wang. "Exploring Causal Information Bottleneck for Adversarial Defense." IEEE Transactions on Information Forensics and Security (2025).
Download Paper | Download Slides | Download Bibtex

AMRM-Pure: Semantic-Preserving Adversarial Purification

Published in AISTATS, 2026

A spotlight paper on adversarial purification published in top-tier conference.

Recommended citation: Zhihao Dou, Zhiqiang Gao, Dongfei Cui, Weida Wang, Qinjian Zhao, Dinggen Zhang, Jun Yan, Zeke Xie, Shufei Zhang. "AMRM-Pure: Semantic-Preserving Adversarial Purification." In AISTATS. 2026.
Download Paper

talks

Talk on Humanoids

Published:

This is a description of your conference proceedings talk, note the different field in type. You can put anything in this field.

Talk on SOTIF of Autonomous Driving

Published:

This is a description of your conference proceedings talk, note the different field in type. You can put anything in this field.

Talk on Adversarial Robustness

Published:

This is a description of your conference proceedings talk, note the different field in type. You can put anything in this field.

teaching

Introduction to AI

Undergraduate course, Information College, Shanghai Ocean University, 2026

It is a course for statistical learning, deep learning, and reinforcement learning