Rihui Sun

Rihui Sun

Auto-generated profile, based on publicly available information

Rihui Sun

Harbin Institute of Technology

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
15 November 2023 in ACM Transactions on Design Automation of Electronic Systems
Verified
Journal Article
Open access
RSRihui Sun
Rihui Sun
PQPengfei Qiu
Pengfei Qiu
YLYongqiang Lyu
Yongqiang Lyu
JDJian Dong
Jian Dong
HWHaixia Wang
Haixia Wang
GQGang Qu
Gang Qu
Graphics Processing Units (GPU) are widely used as deep learning accelerators because of its high performance and low power consumption. Additionally, it remains secure against hardware-induced transient fault injection attacks, a classic type of attacks that have been developed on other computing platforms. In this work, we demonstrate that well-trained machine learning models are robust against hardware fault injection attacks when the faults are generated randomly. However, we discover that these models have components, which we refer to as sensitive targets, that are vulnerable to faults. By exploiting this vulnerability, we propose the Lightning attack, which precisely strikes the model’s sensitive targets with hardware-induced transient faults based on the Dynamic Voltage and Frequency Scaling (DVFS). We design a sensitive targets search algorithm to find the most critical processing units of Deep Neural Network (DNN) models determining the inference results, and develop a genetic algorithm to automatically optimize the attack parameters for DVFS to induce faults. Experiments on three commodity Nvidia GPUs for four widely-used DNN models show that the proposed Lightning attack can reduce the inference accuracy by 69.1% on average for non-targeted attacks, and, more interestingly, achieve a success rate of 67.9% for targeted attacks.

1 November 2026 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
RSRihui Sun
Rihui Sun
JWJin Wu
Jin Wu
HLHanyin Liu
Hanyin Liu
ZTZikang Tao
Zikang Tao
JDJian Dong
Jian Dong
GQGang Qu
Gang Qu
The Global History Register (GHR), an essential component of the Conditional Branch Predictor (CBP) in modern processors, has become a popular attack target because it stores recent branch histories including sensitive kernel control flow information. State-of-the-art GHR attacks can successfully extract bits stored in the GHR, inferring branch directions of executed conditional branch instructions, with conditions linked to secrets. However, retrieving GHR bits requires knowledge of the victim’s virtual address layout. Advancements in enhanced kernel address isolation techniques have significantly reduced the risk of these GHR attacks. In this paper, we introduce GHLeaker, a novel control-flow-based side-channel attack (SCA) that does not rely on kernel address knowledge. GHLeaker retrieves GHR branch information by correlating different branches, effectively bypassing kernel address isolation. We applied GHLeaker to attack Intel, AMD, and ARM processors, demonstrating that all these mainstream processors are vulnerable to the attack. On these platforms, GHLeaker successfully extracts secrets from the Elliptic Curve Digital Signature Algorithm (ECDSA) and Montgomery ladder algorithm which is widely used in cryptography systems such as RSA.

26 July 2026
Verified
Conference Paper
Open access
HLHanyin Liu
Hanyin Liu
JWJin Wu
Jin Wu
RSRihui Sun
Rihui Sun
KRKaiyuan Rong
Kaiyuan Rong
HZHongpei Zheng
Hongpei Zheng
DWDongsheng Wang
Dongsheng Wang