Machine learning attacks on 65nm Arbiter PUFs: Accurate modeling poses strict bounds on usability
The DOI matches its Crossref record; 11 of the 11 references it cites were found in a registry or an index.
Abstract
Arbiter Physically Unclonable Functions (PUFs) have been proposed as efficient hardware security primitives for generating device-unique authentication responses and cryptographic keys. However, the assumed possibility of modeling their underlying challenge-response behavior causes uncertainty about their actual applicability. In this work, we apply well-known machine learning techniques on challenge-response pairs (CRPs) from 64-stage Arbiter PUFs realized in 65nm CMOS, in order to evaluate the effectiveness of such modeling attacks on a modern silicon implementation. We show that a 90%-accurate model can be built from a training set of merely 500 CRPs, and that 5000 CRPs are sufficient to perfectly model the PUFs. To study the implications of these attacks, there is need for a new methodology to assess the security of PUFs suffering from modeling. We propose such a methodology and apply it to our machine learning results, yielding strict bounds on the usability of Arbiter PUFs. We conclude that plain 64-stage Arbiter PUFs are not secure for challenge-response authentication, and the number of extractable secret key bits is limited to at most 600.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Approximation capabilities of multilayer feedforward networks
Hornik, K. (1991) · Neural Networks · doi:10.1016/0893-6080(91)90009-t · Cited by 6,091 · cited 1×
“A strong motivation to use ANN is given by the Universal Approximation Theorem [9], which in short states that a two-layer feedforward ANN containing a finite number of hidden neurons can approximate any function with arbitrary precision.”
MethodA direct adaptive method for faster backpropagation learning: the RPROP algorithm
Riedmiller, M., & Braun, H. (2002) · IEEE International Conference on Neural Networks · doi:10.1109/icnn.1993.298623 · Cited by 3,917 · cited 1×
“In this work, the multilayer ANNs are trained by the resilient backpropagation (RProp) [11] training algorithm because it offers fast convergence .”
MethodPhysical unclonable functions for device authentication and secret key generation
Suh, G. E., & Devadas, S. (2007) · Proceedings of the 44th annual conference on Design automation - DAC '07 · doi:10.1145/1278480.1278484 · Cited by 2,065 · cited 1×
“Their basic idea is to disturb the linearity of the delay model by adding non-linear elements to the response generation, such as feedforward challenge bits [3] and exclusive-or (XOR) combinations of responses [6], [7].”
BackgroundPappu, R., Recht, B., Taylor, J., & Gershenfeld, N. (2002) · Science · doi:10.1126/science.1074376 · Cited by 2,018 · cited 1×
“We first consider the implications of modeling attacks on a PUF-based CRA scheme [12].”
Secret key agreement by public discussion from common information
Maurer, U. M. (1993) · IEEE Transactions on Information Theory · doi:10.1109/18.256484 · Cited by 2,005 · cited 1×
“number of secure key bits that can be extracted from X [14], and it is shown that S(X) = I(X;X ′) with X and X ′ two noisy realisations of the same fuzzy secret.”
BackgroundSilicon physical random functions
Gassend, B., Clarke, D., van Dijk, M., & Devadas, S. (2002) · Proceedings of the 9th ACM conference on Computer and communications security · doi:10.1145/586110.586132 · Cited by 1,660
Modeling attacks on physical unclonable functions
Rührmair, U., Sehnke, F., Sölter, J., Dror, G., Devadas, S., & Schmidhuber, J. (2010) · Proceedings of the 17th ACM conference on Computer and communications security · doi:10.1145/1866307.1866335 · Cited by 1,021 · cited 3×
“Nonetheless, it was demonstrated based on simulations of PUFs [5] that advanced machine learning techniques are still able to generate good models after training with a larger number of CRPs.”
BackgroundMethodLee, J. W., Lim, D., Gassend, B., Suh, G. E., van Dijk, M., & Devadas, S. (2004) · 2004 Symposium on VLSI Circuits. Digest of Technical Papers (IEEE Cat. No.04CH37525) · doi:10.1109/vlsic.2004.1346548 · Cited by 842
Physically Unclonable Functions: A Study on the State of the Art and Future Research Directions
Maes, R., & Verbauwhede, I. (2010) · Information Security and Cryptography · doi:10.1007/978-3-642-14452-3_1 · Cited by 469 · cited 1×
“We refer to [2] for an extensive overview.”
BackgroundSelf-citationMajzoobi, M., Koushanfar, F., & Potkonjak, M. (2008) · 2008 IEEE/ACM International Conference on Computer-Aided Design · doi:10.1109/iccad.2008.4681648 · Cited by 273 · cited 1×
“Besides the simple Arbiter PUF as described above, we will consider the k-XOR Arbiter PUF [5]–[7] consisting of k equally challenged simple Arbiter PUFs, with the response bit of the k-XOR Arbiter PUF being the XOR of the separate k arbiter outputs.”
Method
Cited by
PUF Modeling Attacks on Simulated and Silicon Data
Ruhrmair, U., Solter, J., Sehnke, F., Xu, X., Mahmoud, A., Stoyanova, V., Dror, G., Schmidhuber, J., Burleson, W., & Devadas, S. (2013) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2013.2279798 · Cited by 637
“Please note that both eavesdropping on responses as well as physical access to the PUF is part of the established, general attackmodel for PUFs.”
BackgroundA Lockdown Technique to Prevent Machine Learning on PUFs for Lightweight Authentication
Yu, M.-D., Hiller, M., Delvaux, J., Sowell, R., Devadas, S., & Verbauwhede, I. (2016) · IEEE Transactions on Multi-Scale Computing Systems · doi:10.1109/tmscs.2016.2553027 · Cited by 191
Self-citationSide channel modeling attacks on 65nm arbiter PUFs exploiting CMOS device noise
Delvaux, J., & Verbauwhede, I. (2013) · 2013 IEEE International Symposium on Hardware-Oriented Security and Trust (HOST) · doi:10.1109/hst.2013.6581579 · Cited by 170
“ML techniques require less than 5000 PUF evaluations and are hence faster, as confirmed by measurements on the same chip [2].”
BackgroundMethodSelf-citation
Metadata
- Authors
- Hospodar, G., Maes, R., & Verbauwhede, I.
- Venue
- 2012 IEEE International Workshop on Information Forensics and Security (WIFS)
- Publisher
- IEEE
- Published
- 2012-12
- Pages
- 37-42
- ISSN
- 2157-4774 (ISSN-L)
- OpenAlex
- W2534919326
- Cited by
- 173
- References
- 11
- FWCI
- 5.74
- Type
- conference paper
- Language
- en
- Open access
- closed
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Machine learning attacks on 65nm Arbiter PUFs: Accurate modeling poses strict bounds on usability. https://app.cochat.ai/papers/10.1109/wifs.2012.6412622
- The paper
- Hospodar, G., Maes, R., & Verbauwhede, I. (2012). Machine learning attacks on 65nm Arbiter PUFs: Accurate modeling poses strict bounds on usability. 2012 IEEE International Workshop on Information Forensics and Security (WIFS). https://doi.org/10.1109/wifs.2012.6412622
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.