HH
Hanxun Huang
cs.AIcs.CVcs.LGcs.CRcs.CLstat.MLcs.CYcs.SDcs.IRcs.RO
On Valency
published · living versionsW_8hf6h7cw·v1 · currentpublished
SEEK: Skill Extraction with Evolving Knowledge against Frontier Science Agents
with Cheng Gui, Xiang Zheng, Yutao Wu, Yige Li +3
1 version
Preprints & journals
33 papers in the corpus · 2020–2026RedDarwin: Evolving Jailbreak Skills against Frontier LLMs2610.23726v1 · Zhouyuan Ma, Yutao Wu, Hanxun Huang et al.2026 · 0 citationsarXiv
VEX-Bench: Benchmarking Verification Complexity of LLM-Generated Misinformation2609.35028v1 · Hanxun Huang, Yutao Wu, Qizhou Wang et al.2026 · 0 citationsarXiv
VisER: Visual Evidence and Reliance for Object Hallucination Detection in LVLMs2608.30480v1 · Afsaneh Hasanebrahimi, Hanxun Huang, Christopher Leckie et al.2026 · 0 citationsarXiv
Adversarial Attacks for Good: A Survey of Proactive Protection across the Visual Content Lifecycle2608.04314v1 · Jiaming Zhang, Boyang Chen, Zherui Li et al.2026 · 0 citationsarXiv
GeoDetect: Geometric Adversarial Detection for VLPs2607.14737v1 · Afsaneh Hasanebrahimi, Hanxun Huang, Christopher Leckie et al.2026 · 0 citationsarXiv
HarmProfile: Characterizing Harmful Distributions in Frontier LLMs2608.14577v1 · Zhouyuan Ma, Yutao Wu, Hanxun Huang et al.2026 · 0 citationsarXiv
Density-Aware Translation of Spurious Correlations in Zero-Shot VLMs2606.01710v1 · Afsaneh Hasanebrahimi, Hanxun Huang, Christopher Leckie et al.2026 · 0 citationsarXiv
When Medical Safety Alignment Fails: A Benchmark for Evaluating LLMs on High-Risk Medical Queries2606.28332v1 · Yige Li, Jun Sun, Wei Zhao et al.2026 · 0 citationsarXiv
AudioMosaic: Contrastive Masked Audio Representation Learning2605.14231v1 · Hanxun Huang, Qizhou Wang, Xingjun Ma et al.2026 · 0 citationsarXiv
Safety in Embodied AI: A Survey of Risks, Attacks, and Defenses2605.02900v2 · Xiao Li, Xiang Zheng, Yifeng Gao et al.2026 · 0 citationsarXiv
Backdoor4Good: Benchmarking Beneficial Uses of Backdoors in LLMs2603.07452v1 · Yige Li, Wei Zhao, Zhe Li et al.2026 · 0 citationsarXiv
Internal Safety Collapse in Frontier Large Language Models2603.23509v1 · Yutao Wu, Xiao Liu, Yifeng Gao et al.2026 · 0 citationsarXiv
Semantic-aware Adversarial Fine-tuning for CLIP2602.12461v1 · Jiacheng Zhang, Jinhao Li, Hanxun Huang et al.2026 · 0 citationsarXiv
Toward Universal and Transferable Jailbreak Attacks on Vision-Language Models2602.01025v1 · Kaiyuan Cui, Yige Li, Yutao Wu et al.2026 · 0 citationsarXiv
Just Ask: Curious Code Agents Reveal System Prompts in Frontier LLMs2601.21233v2 · Xiang Zheng, Yutao Wu, Hanxun Huang et al.2026 · 0 citationsarXiv
AutoBackdoor: Automating Backdoor Attacks via LLM Agents2511.16709v1 · Yige Li, Zhe Li, Wei Zhao et al.2025 · 0 citationsarXiv
AUDETER: A Large-scale Dataset for Deepfake Audio Detection in Open Worlds2509.04345v2 · Qizhou Wang, Hanxun Huang, Guansong Pang et al.2025 · 1 citationarXiv
T2UE: Generating Unlearnable Examples from Text Descriptions2508.03091v1 · Xingjun Ma, Hanxun Huang, Tianwei Song et al.2025 · 1 citationarXiv
X-Transfer Attacks: Towards Super Transferable Adversarial Attacks on CLIP2505.05528v3 · Hanxun Huang, Sarah Erfani, Yige Li et al.2025 · 0 citationsarXiv
Geometry-Guided Adversarial Prompt Detection via Curvature and Local Intrinsic Dimension2503.03502v2 · Canaan Yung, Hanxun Huang, Christopher Leckie et al.2025 · 0 citationsarXiv
Detecting Backdoor Samples in Contrastive Language Image Pretraining2502.01385v2 · Hanxun Huang, Sarah Erfani, Yige Li et al.2025 · 0 citationsarXiv
Towards Million-Scale Adversarial Robustness Evaluation With Stronger Individual Attacks2411.15210v4 · Yong Xie, Weijie Zheng, Hanxun Huang et al.2024 · 0 citationsarXiv
Expose Before You Defend: Unifying and Enhancing Backdoor Defenses via Exposed Models2410.19427v1 · Yige Li, Hanxun Huang, Jiaming Zhang et al.2024 · 0 citationsarXiv
BackdoorLLM: A Comprehensive Benchmark for Backdoor Attacks and Defenses on Large Language Models2408.12798v2 · Yige Li, Hanxun Huang, Yunhan Zhao et al.2024 · 8 citationsarXiv
Downstream Transfer Attack: Adversarial Attacks on Downstream Models with Pre-trained Vision Transformers2408.01705v1 · Weijie Zheng, Xingjun Ma, Hanxun Huang et al.2024 · 1 citationarXiv
Shortcuts Everywhere and Nowhere: Exploring Multi-Trigger Backdoor Attacks2401.15295v4 · Yige Li, Jiabo He, Hanxun Huang et al.2024 · 6 citationsarXiv
LDReg: Local Dimensionality Regularized Self-Supervised Learning2401.10474v2 · Hanxun Huang, Ricardo J. G. B. Campello, Sarah Monazam Erfani et al.2024 · 0 citationsarXiv
Distilling Cognitive Backdoor Patterns within an Image2301.10908v4 · Hanxun Huang, Xingjun Ma, Sarah Erfani et al.2023 · 13 citationsarXiv
Exploring Architectural Ingredients of Adversarially Robust Deep Neural Networks2110.03825v5 · Hanxun Huang, Yisen Wang, Sarah Monazam Erfani et al.2021 · 44 citationsarXiv
Unlearnable Examples: Making Personal Data Unexploitable2101.04898v2 · Hanxun Huang, Xingjun Ma, Sarah Monazam Erfani et al.2021 · 68 citationsarXiv
Neural Architecture Search via Combinatorial Multi-Armed Bandit2101.00336v2 · Hanxun Huang, Xingjun Ma, Sarah M. Erfani et al.2021 · 7 citationsInternational Joint Conference on Neural Networks (IJCNN) 2021
Normalized Loss Functions for Deep Learning with Noisy Labels2006.13554v1 · Xingjun Ma, Hanxun Huang, Yisen Wang et al.2020 · 314 citationsarXiv
Imbalanced Gradients: A Subtle Cause of Overestimated Adversarial Robustness2006.13726v4 · Xingjun Ma, Linxi Jiang, Hanxun Huang et al.2020 · 7 citationsarXiv
Career total: 39 works. 33 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.