Improving Machine Learning Attack Resiliency via Conductance Balancing in Memristive Strong PUFs

Larimian, S.; Mahmoodi, Mohammad Reza; Strukov, Dmitri B.

doi:10.1109/ted.2022.3152469

Cited by 4 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The input value is used as the coefficient or weight of the function to obtain the function output value or predicted value. The relationship between the input independent variable x and the function prediction value f(x) is [20]:…”

Section: Modeling Attack Algorithmsmentioning

confidence: 99%

A highly stable XOR APUF based on deviation signal screening mechanism

Zhai,

Wang,

et al. 2023

IEICE Electron. Express

View full text Add to dashboard Cite

By using the XOR method to obfuscate response of Arbiter PUFs (APUFs), the ability to resist machine learning (ML) modeling attacks can be improved, but this method will reduce the stable of PUF to a certain extent, making PUF lose usability. In view of this, a highly stable XOR APUF (HS-XOR APUF) circuit is proposed by studying the generation mechanism of unstable response bits and the screening characteristics of the logic gates delay signals. First, the maximum and minimum delay signals are screened with AND/OR gate to generate highly stable response bits; Then, by inserting a inverter in the APUF switch unit, the delay time of the signal is increased, and the influence of environmental factors on the comparison signal with small delay deviation is weakened; Finally, the FPGA experimental results show that the response stability of HS-XOR APUF is rarely affected by the count of XOR units, and the AND/OR gate delay signal screening structure is simple and consumes a small amount of hardware resources, which can be widely used in the fields of information security.

show abstract

Section: Modeling Attack Algorithmsmentioning

confidence: 99%

A highly stable XOR APUF based on deviation signal screening mechanism

Zhai,

Wang,

et al. 2023

IEICE Electron. Express

View full text Add to dashboard Cite

show abstract

“…In the study [85], It is shown how to create a highly secure neuromorphic system utilizing a physically unclonable function (PUF) that makes use of high entropy produced by a memristor's stochastically switching made of poly (1,3,5trivinyl-1,3,5-trimethyl cyclotrisiloxane) (pV3D3). The authors [86] suggest two methods for enhancing PUFs' resistance to ML attacks. Each cross-point device should contribute as much as possible to the PUF output in order to reduce the predictability of the response, which is the general concept behind both ideas.…”

Section: Implementation Of Memristive Puf (M-puf)mentioning

confidence: 99%

Memristive Physical Unclonable Functions: The State-of-the-Art Technology

Al-Khaboori,

Al-Mashhadani

2023

IJSSE

View full text Add to dashboard Cite

Internet of Things connected many useful electronic devices to each other through the internet, and sharing private and sensitive data between these devices needs secure access and communication. One of the best solutions for this purpose is hardware security primitives such as Physically Unclonable Functions (PUFs). PUFs are cryptographic primitives that are employed to produce a unique and reliable digital fingerprint for a particular electronic circuit. This digital fingerprint is used in many security applications such as chip identification, authentication, and secret key storage and generation. The emergence of memristors (Memory-Resistor) as new nanotechnologies are utilized extensively in hardware security applications such as Memristive PUFs. Research progress in Memristive PUFs resulted in improved performance metrics of PUFs due to memristors' unique characteristics. This article provides an investigation of different design approaches of Memristive PUFs that were introduced in the literature. Then, provide detailed performance evaluation results obtained by simulation and fabrication processes for different Memristive PUFs designs, and make a comparison between these results. Finally, concluded that most of the circuits are evaluated by simulation, whereas few other circuits were evaluated by fabrication owing to the expensive fabrication process. Since the memristor is a prototype and not commercialized yet, it is expected to be adopted and marketed in the next generation of hardware security.

show abstract

“…Still, practical numerical problems often require solutions with single (223≈107levels, or ~10 −7 error) or double precision (252≈1015 levels, or ~10 −15 error). Accordingly, analog devices have been primarily used for applications without high-precision requirements, such as machine learning ( 8 , 27 – 31 ), randomness-based processing like stochastic computing ( 32 – 34 ), and hardware security ( 35 – 37 ). To achieve high-precision solutions, innovations in architecture and algorithms, codesigned with analog devices, must be made.…”

mentioning

confidence: 99%

Programming memristor arrays with arbitrarily high precision for analog computing

Song,

Rao,

et al. 2024

Science

View full text Add to dashboard Cite

In-memory computing represents an effective method for modeling complex physical systems that are typically challenging for conventional computing architectures but has been hindered by issues such as reading noise and writing variability that restrict scalability, accuracy, and precision in high-performance computations. We propose and demonstrate a circuit architecture and programming protocol that converts the analog computing result to digital at the last step and enables low-precision analog devices to perform high-precision computing. We use a weighted sum of multiple devices to represent one number, in which subsequently programmed devices are used to compensate for preceding programming errors. With a memristor system-on-chip, we experimentally demonstrate high-precision solutions for multiple scientific computing tasks while maintaining a substantial power efficiency advantage over conventional digital approaches.

show abstract

Improving Machine Learning Attack Resiliency via Conductance Balancing in Memristive Strong PUFs

Cited by 4 publications

References 14 publications

A highly stable XOR APUF based on deviation signal screening mechanism

A highly stable XOR APUF based on deviation signal screening mechanism

Memristive Physical Unclonable Functions: The State-of-the-Art Technology

Programming memristor arrays with arbitrarily high precision for analog computing

Contact Info

Product

Resources

About