Lihui Pang

Dr. Lihui Pang


Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
2 May 2022 in IEEE Transactions on Emerging Topics in Computing
Verified
Journal Article
Closed access
JYJianrong Yao
Jianrong Yao
LPLihui Pang
Lihui Pang
YSYang Su
Yang Su
ZZZhi Zhang
Zhi Zhang
WYWei Yang
Wei Yang
YGYansong Gao
Yansong Gao
Physical Unclonable Function (PUF) is a hardware security primitive with a desirable feature of low cost. Considering the space of challenge-response pairs (CRPs), there are two PUF categories: weak PUF and strong PUF. Compared to a weak PUF, a strong PUF has a wider range of applications. However, it is challenging to design a reliable and secure lightweight strong PUF. To address this challenge, PUF recomposition built upon multiple simple PUF instances has received a lot of attention in research, such as the most popular XOR-APUF, the recent MPUF in IEEE TC 2017 (Sahoo et al. , 2018), XOR-FF-APUF in IEEE TIFS 2020 (Avvaru et al. , 2020) and IPUF in TCHES 2020 (Nguyen et al. , 2019). When a combination of MAX and MIN (equal to AND and OR) bitwise operations are used in PUF recomposition (Rührmair et al. , 2010), (Rührmair et al. , 2013) and (Gao et al. , 2020), its resilience against model attacks was expected to be improved markedly, because one bitwise operation might be vulnerable to one type of modeling attack and combining them can yield improved resilience. To our knowledge, there is no explicit evaluation of this recomposition; thus, this study is the first to evaluate the uniformity and reliability of the O R- A ND- X OR -PUF (OAX-PUF)—( $x,y,z$ )-OAX-PUF. Compared to the most used $l$ -XOR-PUF, the ( $x,y,z$ )-OAX-PUF shows better reliability given $l=x+y+z$ without degrading the uniformity (i.e., retain to be 50%). As APUF is a compact PUF instance for constructing lightweight strong PUF candidates, e.g., XOR-APUF, MUXPUF and IPUF, we further examine the modeling resilience of the ( $x,y,z$ )-OAX-APUF using four powerful attacks, i.e., logistic regression (LR), reliability assisted CMA-ES, multilayer perceptron (MLP), and the most recent hybrid LR-reliability. Compared to the XOR-APUF, the OAX-APUF successfully defeats the CMA-ES attack. It often shows improved modeling resilience against LR and hybrid LR-reliability attacks while always increasing the attacking time costs of these two attacks. However, OAX-APUF exhibits lower modeling resilience against the MLP attack unless the $x,y,z$ are carefully tuned. Overall, the OAX recomposition could be an alternative lightweight recomposition approach in constructing strong PUFs if the underlying PUF (e.g., FF-APUF), has shown improved resilience against modeling attacks, because the OAX incurs smaller reliability degradation compared to XOR.

12 January 2023 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
YGYansong Gao
Yansong Gao
JYJianrong Yao
Jianrong Yao
LPLihui Pang
Lihui Pang
WYWei Yang
Wei Yang
AFAnmin Fu
Anmin Fu
DADerek Abbott
Derek Abbott
To improve the modeling resilience of silicon strong physical unclonable functions (PUFs), in particular, the APUFs that yield a very large number of challenge response pairs (CRPs), a number of composited APUF variants such as XOR-APUF, interpose-PUF (iPUF), feed-forward APUF (FF-APUF), and OAX-APUF have been devised. When examining their security in terms of modeling resilience, utilizing multiple information sources such as power side channel information (SCI) or/and reliability SCI given a challenge is under-explored, which poses a challenge to their supposed modeling resilience in practice. Building upon multi-label/head deep learning model architecture, this work proposes Multi-Label Multi-Side-channel-information enabled deep learning Attacks (MLMSA) to thoroughly evaluate the modeling resilience of aforementioned APUF variants. Despite its simplicity, MLMSA can successfully break large-scaled APUF variants, which has not previously been achieved. More precisely, the MLMSA breaks 128-stage 30-XOR-APUF, (9,9)-and (2,18)-iPUFs, and (2,2,30)-OAX-APUF when CRPs, power SCI and reliability SCI are concurrently used. It breaks 128-stage 12-XOR-APUF and (2,2,9)-OAX-APUF even when only the easy-to-obtain reliability SCI and CRPs are exploited. The 128-stage six-loop FF-APUF and one-loop 20-XOR-FF-APUF can be broken by simultaneously using reliability SCI and CRPs. All these attacks are normally completed within an hour with a standard personal computer. Therefore, MLMSA is a useful technique for evaluating other existing or any emerging strong PUF designs.

14 December 2022 in EURASIP Journal on Advances in Signal Processing
Verified
Journal Article
Open access
PDF file
LPLihui Pang
Lihui Pang
YTYilong Tang
Yilong Tang
QTQingyi Tan
Qingyi Tan
YLYulang Liu
Yulang Liu
BYBin Yang
Bin Yang
The blind signal separation (BSS) algorithm obtains each original/source signal from the observed signal collected by the receiving antenna or sensor. Objective/loss/cost function and optimization method are two key parts of BSS algorithm. Modifying the objective function and optimization from the perspective of neural network (NN) is a novel concept in BSS domain. $$L_2$$L2 regularization is adopted as a term of maximum likelihood estimation (MLE)-based objective function like in Liu et al. (Sensors 21(3):973, 2021); however, we modified the probability density function (PDF) term of the objective function and used the kernel density estimation method for time–frequency overlapped digital communication signal. Multiple optimizers are studied in this paper, and we figure out the right optimizer for our application scenario. A varies of comparison experiments—whose separation results will be provided in forms of correlation coefficient and performance index—are carried out, which indicate our method can converge quickly and achieve satisfactory separation results with performance index (PI) lower than 0.02 when signal-to-noise ratio (SNR) no less than 10dB. Additionally, it demonstrates performance of our method is better than that of typical separation—FastICA, especially for the lower SNR environment, and it shows that our method is not sensitive to the frequency overlap level (FOL) of the source signal, even FOL as high as $$100\%$$100% ; it still can get high-precision separation results with $$\textrm{PI}<0.02$$PI<0.02 .

26 February 2024 in IEEE Access
Verified
Journal Article
Open access
BHBenxuan Huang
Benxuan Huang
LPLihui Pang
Lihui Pang
AFAnmin Fu
Anmin Fu
SASaid F. Al-Sarawi
Said F. Al-Sarawi
DADerek Abbott
Derek Abbott
YGYansong Gao
Yansong Gao
The motivation for the development of multi-exit networks (MENs) lies in the desire to minimize the delay and energy consumption associated with the inference phase. Moreover, MENs are designed to expedite predictions for easily identifiable inputs by allowing them to exit the network prematurely, thereby reducing the computational burden due to challenging inputs. Nevertheless, there is a lack of comprehensive understanding regarding the security vulnerabilities inherent in MENs. In this study, we introduce a novel approach called the sponge attack , which aims to compromise the fundamental advantages of MENs that allow easily identifiable images to leave in early exits. By employing data poisoning techniques, we frame the sponge attack as an optimization problem that empowers an attacker to select a specific trigger, such as adverse weather conditions (e.g., raining), to compel inputs to traverse the complete network layers of the MEN (e.g., in the context of traffic sign recognition) instead of early-exits when the trigger condition is met. Remarkably, our attack has the capacity to increase inference latency, while maintaining the classification accuracy even in the presence of a trigger, thus operating discreetly. Extensive experimentation on three diverse natural datasets (CIFAR100, GTSRB, and STL10), each trained with three prominent MEN architectures (VGG16, ResNet56, and MSDNet), validates the efficacy of our attack in terms of latency augmentation and its effectiveness in preserving classification accuracy under trigger conditions.

18 January 2025 in Computer Communications
Verified
Journal Article
Closed access
TLTing Luo
Ting Luo
HPHuaibing Peng
Huaibing Peng
AFAnmin Fu
Anmin Fu
WYWei Yang
Wei Yang
LPLihui Pang
Lihui Pang
YGYansong Gao
Yansong Gao

10 September 2024 in IEEE Transactions on Aerospace and Electronic Systems
Verified
Journal Article
Closed access
LPLihui Pang
Lihui Pang
YTYilong Tang
Yilong Tang
KJKaili Jiang
Kaili Jiang
WZWenwei Zhang
Wenwei Zhang
In nowadays information era, the types of communication or radar electronic equipment for both military and civilian applications are increasing significantly, which leads to various communication and radar signals overcrowded in time domain, overlapped in frequency domain and intertwined in space domain shaping a more complicated electromagnetic environment. In order to accurately capture the interested signal or discover the interference signal, separating these time-frequency overlapped signals and extracting the information implied in the useful signal have become a research task with significant for electromagnetic surveillance domain. This paper investigated the sinusoidal similarity among electromagnetic reconnaissance signals and proposed a time-Frequency overlapped signals separation method based on sinusoidal similarity for complex electromagnetic monitoring environments. Firstly, we mathematically derived the sinusoidal similarity among electromagnetic reconnaissance signals for three different mixed scenarios, to be specify, digital modulation communication signals mixing scene (CCMs), digital modulation communication signal(s) and Pulse Repetition Interval Radar signal(s) mixing scene (CRMs), Pulse Repetition Interval Radar signals mixing scene (RRMs). Then, we formulated a blind signal separation objective/cost function utilizing sinusoidal similarity (SSim) among electromagnetic reconnaissance signals. In addition to the SSim constraint, we introduced another constraint term into the objective function, accounting for the unique characteristics of source signals in various mixed scenarios. After that, we successfully optimized the constructed objective function by leveraging advanced machine learning optimization algorithms. Then, we substantiated the validity of the proposed method through simulation experiments, showcasing its performance from various perspectives, and we underscored its advantages by conducting comparisons with classical methods such as FastICA and JADE. Furthermore, the proposed Blind Signal Separation (BSS) method has broader applications in the realm of artificial intelligence. It can be employed for intelligent tasks like speech signal separation and biomedical signal separation, showcasing its versatility in diverse domains.