Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Lihe Zhang

Lihe Zhang

3 indexed papers

Recent (6 mo)
3
With code
0
Influential cites
0
Benchmarked
0

Publications per year

3
26

Top categories

Crypto×3ML×2AI×1

Frequent co-authors

Mengnan Zhao3×
Baocai Yin2×
Tianhang Zheng2×
Bo Wang2×
Hong Zhong1×
Geyong Min1×

Research Timeline

2026
Unveiling the Backdoor Mechanism Hidden Behind Catastrophic Overfitting in Fast Adversarial Training

This paper reinterprets catastrophic overfitting (CO) in Fast Adversarial Training (FAT) as a weak backdoor mechanism, proposing backdoor-inspired strategies to mitigate this generalization failure.

Mitigating Error Amplification in Fast Adversarial Training

The paper proposes a Distribution-aware Dynamic Guidance (DDG) strategy to mitigate catastrophic overfitting and the robustness-accuracy trade-off inherent in Fast Adversarial Training (FAT) by dynamically adjusting perturbation budgets and supervision signals based on sample confidence.

CoreUnlearn: Rethinking Concept Unlearning through Disentangled Component-Level Erasure in Text-guided Diffusion Models

CoreUnlearn introduces a novel framework that disentangles and removes undesirable concepts from text-guided diffusion models by targeting specific, erasure-critical components of the concept embedding, thereby achieving effective unlearning with minimal performance degradation.

Highlighted terms show continued research focus across papers

Papers

cs.CRRecentJun 1, 2026

CoreUnlearn: Rethinking Concept Unlearning through Disentangled Component-Level Erasure in Text-guided Diffusion Models

Mengnan Zhao, Lihe Zhang, Baocai Yin

CoreUnlearn introduces a novel framework that disentangles and removes undesirable concepts from text-guided diffusion models by targeting specific, erasure-critical components of the concept embeddin…

View →
cs.LGcs.AIcs.CRRecentApr 27, 2026

Unveiling the Backdoor Mechanism Hidden Behind Catastrophic Overfitting in Fast Adversarial Training

Mengnan Zhao, Lihe Zhang, Tianhang Zheng, Bo Wang +1 more

This paper reinterprets catastrophic overfitting (CO) in Fast Adversarial Training (FAT) as a weak backdoor mechanism, proposing backdoor-inspired strategies to mitigate this generalization failure.

View →
cs.LGcs.CRRecentApr 27, 2026

Mitigating Error Amplification in Fast Adversarial Training

Mengnan Zhao, Lihe Zhang, Bo Wang, Tianhang Zheng +2 more

The paper proposes a Distribution-aware Dynamic Guidance (DDG) strategy to mitigate catastrophic overfitting and the robustness-accuracy trade-off inherent in Fast Adversarial Training (FAT) by dynami…

View →