News

Relational Feature Distillation for Lightweight 3D Point Cloud Segmentation was accepted to NeurIPS 2026.

On the Vulnerability of Lane Detectors to Physical Shadow Attack was accepted to ACSAC 2026.

GATE was accepted to BMVC 2026.

Distilling Vision-Language Models for Robust Traffic Sign Perception in Autonomous Vehicles was accepted to IROS 2026.

Budget-Aware Adaptive Adversarial Patches for Black-Box Object Detection was accepted to IEEE ICIP 2026.

From MIRAGE to CLEAR was accepted to IEEE/IFIP DSN 2026.

Three papers were accepted to the CVPR 2026 Workshops / 6th AdvML Workshop.

• Interpretable Adversarial Prompt Tuning via Semantic Concepts
• SASA: Sequence-Aware Shadow Attacks via Attention Alignment for Traffic Sign Recognition
• Auditing Traffic-Sign Robustness via DDIM Inversion: Do Diffusion Latents Preserve Shadow Attacks?

SLNet was accepted to ICRA 2026.

Three papers were accepted to the IEEE Intelligent Vehicles Symposium 2026.

• NPNet: A Non-Parametric Network with Adaptive Gaussian-Fourier Positional Encoding for 3D Classification and Segmentation
• Comparative Analysis of Patch Attack on VLM-Based Autonomous Driving Architectures
• Toward Inherently Robust VLMs Against Visual Perception Attacks

Avoiding the Crash received the Best Paper Award at SAE/WCX 2025.

Serving as key personnel on the U.S. DOT UTC TraCR project Resilient Autonomous Vehicle Perception Under Adversarial Settings.