Guanhong Tao

发表

Attacks Meet Interpretability: Attribute-steered Detection of Adversarial Samples pdf

Xiangyu Zhang, Guanhong Tao, Yingqi Liu, 2018, NeurIPS.

NIC: Detecting Adversarial Samples with Neural Network Invariant Checking

Wen-Chuan Lee, Xiangyu Zhang, Guanhong Tao, 2019, NDSS.

Physical Attack on Monocular Depth Estimation with Optimal Adversarial Patches

Guanhong Tao, Hongjun Choi, Zhiyuan Cheng, 2022, ECCV.

Piccolo: Exposing Complex Backdoors in NLP Transformer Models

X. Zhang, Shiqing Ma, Guanhong Tao, 2022, 2022 IEEE Symposium on Security and Privacy (SP).

FLIP: A Provable Defense Framework for Backdoor Mitigation in Federated Learning

Pin-Yu Chen, Shiqing Ma, Guanhong Tao, 2022, ArXiv.

RULER: discriminative and iterative adversarial training for deep neural network fairness

Guanhong Tao, Chunrong Fang, Tingxu Han, 2022, ESEC/SIGSOFT FSE.

BEAGLE: Forensics of Deep Learning Backdoor Attack for Better Defense

Shiqing Ma, Guanhong Tao, Siyuan Cheng, 2023, NDSS.

MIRROR: Model Inversion for Deep LearningNetwork with High Fidelity

X. Zhang, Guanhong Tao, Yuan Yao, 2022, NDSS.

ABS: Scanning Neural Networks for Back-doors by Artificial Brain Stimulation

Xiangyu Zhang, Guanhong Tao, Wen-Chuan Lee, 2019, CCS.

CPC: Automatically Classifying and Propagating Natural Language Comments via Program Analysis

Weifeng Zhang, Xiangyu Zhang, Yu Shi, 2020, 2020 IEEE/ACM 42nd International Conference on Software Engineering (ICSE).

Complex Backdoor Detection by Symmetric Feature Differencing

X. Zhang, Shiqing Ma, Yingqi Liu, 2022, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Constrained Optimization with Dynamic Bound-scaling for Effective NLPBackdoor Defense

X. Zhang, Shiqing Ma, Yingqi Liu, 2022, ICML.

Better Trigger Inversion Optimization in Backdoor Scanning

X. Zhang, Shiqing Ma, Yingqi Liu, 2022, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Backdoor Scanning for Deep Neural Networks through K-Arm Optimization

X. Zhang, Shiqing Ma, Guanhong Tao, 2021, ICML.

Correlations between deep neural network model coverage criteria and model quality

Xiangyu Zhang, Lei Xu, Shiqing Ma, 2020, ESEC/SIGSOFT FSE.

Fusion is Not Enough: Single-Modal Attacks to Compromise Fusion Models in Autonomous Driving

Guanhong Tao, Hongjun Choi, Michael Zuzak, 2023, ArXiv.

Gradient Shaping: Enhancing Backdoor Attack Against Reverse Engineering

Shiqing Ma, Haixu Tang, Guanhong Tao, 2023, ArXiv.

Adversarial Training of Self-supervised Monocular Depth Estimation against Physical-World Attacks

Guanhong Tao, Zhiyuan Cheng, James Liang, 2023, ICLR.

ALchemist: Fusing Application and Audit Logs for Precise Attack Provenance without Instrumentation

Vinod Yegneswaran, Xiangyu Zhang, Dongyan Xu, 2021, NDSS.

Backdoor Vulnerabilities in Normally Trained Deep Learning Models

Shiqing Ma, Guanhong Tao, Zhuo Zhang, 2022, ArXiv.

Backdooring Neural Code Search

Guanhong Tao, Chunrong Fang, Xiangyu Zhang, 2023, 2305.17506.

DECK: Model Hardening for Defending Pervasive Backdoors

Guanhong Tao, Zhuo Zhang, Siyuan Cheng, 2022, ArXiv.

Model Orthogonalization: Class Distance Hardening in Neural Networks for Better Security

X. Zhang, Guanhong Tao, Zhuo Zhang, 2022, 2022 IEEE Symposium on Security and Privacy (SP).

Bounded Adversarial Attack on Deep Content Features

Guanhong Tao, Xiangyu Zhang, Qiuling Xu, 2022, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

StochFuzz: Sound and Cost-effective Fuzzing of Stripped Binaries by Incremental and Stochastic Rewriting

Wei You, Guanhong Tao, Yousra Aafer, 2021, 2021 IEEE Symposium on Security and Privacy (SP).

Detecting Backdoors in Pre-trained Encoders

Shiqing Ma, Guanhong Tao, Siyuan Cheng, 2023, ArXiv.

FLIP: A Provable Defense Framework for Backdoor Mitigation in Federated Learning

Pin-Yu Chen, Shiqing Ma, Guanhong Tao, 2022, ICLR.

Deep Learning for Additive Screening in Perovskite Light-Emitting Diodes.

L. Zhang, Guanhong Tao, Chen Cao, 2022, Angewandte Chemie.

Towards Feature Space Adversarial Attack by Style Perturbation

Siyuan Cheng, Guanhong Tao, Qiuling Xu, 2021, AAAI.

Precise Android API Protection Mapping Derivation and Reasoning

Ninghui Li, Xiangyu Zhang, Jianjun Huang, 2018, CCS.

BDA: practical dependence analysis for binary executables by unbiased whole-program path sampling and per-path abstract interpretation

Wei You, Xiangyu Zhang, Zhuo Zhang, 2019, Proc. ACM Program. Lang..

OSPREY: Recovery of Variable and Data Structure via Probabilistic Analysis for Stripped Binary

Wei You, Yonghwi Kwon, Guanhong Tao, 2021, 2021 IEEE Symposium on Security and Privacy (SP).

Improving Binary Code Similarity Transformer Models by Semantics-Driven Instruction Deemphasis

Guanhong Tao, Qingkai Shi, Yapeng Ye, 2023, ISSTA.

FIRM: DETECTING ADVERSARIAL AUDIOS

Yunhan Jia, Shiqing Ma, Guanhong Tao, 2021 .

Towards Feature Space Adversarial Attack pdf

Xiangyu Zhang, Siyuan Cheng, Guanhong Tao, 2020, ArXiv.

TRADER: Trace Divergence Analysis and Embedding Regulation for Debugging Recurrent Neural Networks

Xiangyu Zhang, Yingqi Liu, Shiqing Ma, 2020, 2020 IEEE/ACM 42nd International Conference on Software Engineering (ICSE).

PELICAN: Exploiting Backdoors of Naturally Trained Deep Learning Models In Binary Code Analysis

Guanhong Tao, Yapeng Ye, Xiangyu Zhang, 2023, USENIX Security Symposium.

Hard-label Black-box Universal Adversarial Patch Attack

Guanhong Tao, Siyuan Cheng, Xiangyu Zhang, 2023, USENIX Security Symposium.

Backdooring Neural Code Search

Guanhong Tao, Chunrong Fang, Quanjun Zhang, 2023, ACL.

ImU: Physical Impersonating Attack for Face Recognition System with Natural Style Changes

Y. Yao, Shiqing Ma, Guanhong Tao, 2023, 2023 IEEE Symposium on Security and Privacy (SP).

Distribution Preserving Backdoor Attack in Self-supervised Learning

Shiqing Ma, Guanhong Tao, Zhenting Wang, 2022 .

Make Them Spill the Beans! Coercive Knowledge Extraction from (Production) LLMs

Guanhong Tao, Siyuan Cheng, Xiangyu Zhang, 2023, 2312.04782.

Elijah: Eliminating Backdoors Injected in Diffusion Models via Distribution Shift

Guanhong Tao, Tsung-Yi Ho, Siyuan Cheng, 2023, AAAI.

PEM: Representing Binary Program Semantics for Similarity Analysis via a Probabilistic Execution Model

Guanhong Tao, Qingkai Shi, Yapeng Ye, 2023, ESEC/SIGSOFT FSE.

Rapid Optimization for Jailbreaking LLMs via Subconscious Exploitation and Echopraxia

Guanhong Tao, Siyuan Cheng, Shengwei An, 2024, ArXiv.

This paper is included in the Proceedings of the 32nd USENIX Security Symposium.

Guanhong Tao, Siyuan Cheng, Xiangyu Zhang, 2022 .

D-square-B: Deep Distribution Bound for Natural-looking Adversarial Attack pdf

Xiangyu Zhang, Guanhong Tao, Qiuling Xu, 2020, ArXiv.

Fusion is Not Enough: Single Modal Attacks on Fusion Models for 3D Object Detection

Guanhong Tao, Michael Zuzak, Zhiyuan Cheng, 2023, 2304.14614.