Publications

2023

• Synthèse de champs nuageux sous contraintes physiques
  
  • Chatillon Pierrick
  • Gousseau Yann
  • Lefebvre Sidonie
  , 2023. Nous proposons dans cet article une architecture de réseau de neurone et une forme de supervision permettant de spécifier des paramètres physiques en entrée d'un réseau générateur et d'obtenir en sortie des images texturées de nuages respectant les statistiques correspondantes. Ce modèle paramétrique contrôle la pente spectrale et la distribution des niveaux de gris des images. La capacité du modèle à générer des images sous contraintes est testée sur des images de télédétection dans l'infrarouge et des images de contenu en eau simulées.
• Study of historical Byzantine seal images: the BHAI project for computer-based sigillography
  
  • Eyharabide Victoria
  • Likforman-Sulem Laurence
  • Orlandi Lucia Maria
  • Binoux Alexandre
  • Rageau Theophile
  • Huang Qijia
  • Fiandrotti Attilio
  • Caseau Béatrice
  • Bloch Isabelle
  , 2023, pp.49-54. BHAI (Byzantine Hybrid Artificial Intelligence) is the first project based on artificial intelligence dedicated to Byzantine seals. The scientific consortium comprises a multidisciplinary team involving historians specialized in the Byzantine period, specialists of sigillography, and computer science experts. This article describes the main objectives of this project: data acquisition of seal images, text and iconography recognition, seal dating, as well as our current achievements. (10.1145/3604951.3605523)
  DOI : 10.1145/3604951.3605523
• RaSP: Relation-aware Semantic Prior for Weakly Supervised Incremental Segmentation
  
  • Roy Subhankar
  • Volpi Riccardo
  • Csurka Gabriela
  • Larlus Diane
  , 2023. Class-incremental semantic image segmentation assumes multiple model updates, each enriching the model to segment new categories. This is typically carried out by providing expensive pixel-level annotations to the training algorithm for all new objects, limiting the adoption of such methods in practical applications. Approaches that solely require image-level labels offer an attractive alternative, yet, such coarse annotations lack precise information about the location and boundary of the new objects. In this paper we argue that, since classes represent not just indices but semantic entities, the conceptual relationships between them can provide valuable information that should be leveraged. We propose a weakly supervised approach that exploits such semantic relations to transfer objectness prior from the previously learned classes into the new ones, complementing the supervisory signal from image-level labels. We validate our approach on a number of continual learning tasks, and show how even a simple pairwise interaction between classes can significantly improve the segmentation mask quality of both old and new classes. We show these conclusions still hold for longer and, hence, more realistic sequences of tasks and for a challenging few-shot scenario.
• Teleportation-Based Error Correction Protocol of Time–Frequency Qubit States
  
  • Fabre Nicolas
  Applied Sciences, Multidisciplinary digital publishing institute (MDPI), 2023, 13 (16), pp.9462. We present a linear optical protocol for teleporting and correcting both temporal and frequency errors in two time–frequency qubit states. The first state is the frequency (or time-of-arrival) cat qubit, which is a single photon in a superposition of two frequencies (or time-of-arrival), while the second is the time–frequency Gottesman–Kitaev–Preskill (GKP) state, which is a single photon with a frequency comb structure. The proposed optical scheme could be valuable for reducing the error rate in quantum communication protocols involving one of these qubits. (10.3390/app13169462)
  DOI : 10.3390/app13169462
• Capacity ATL
  
  • Ballot Gabriel
  • Malvone Vadim
  • Leneutre Jean
  • Laarouchi Youssef
  , 2023. Model checking strategic abilities was successfully developed and applied since the early 2000s to ensure properties in Multi-Agent System. In this paper, we introduce the notion of capacities giving different abilities to an agent. This applies naturally to systems where multiple entities can play the same role in the game, such as different client versions in protocol analysis, different robots in heterogeneous fleets, different personality traits in social structure modeling, or different attacker profiles in a cybersecurity setting. With the capacity of other agents being unknown at the beginning of the game, the longstanding problems of imperfect information arise. Our contribution is the following: (i) we define a new class of concurrent game structures where the agents have different capacities that modify their action list and (ii) we introduce a logic extending Alternating-time Temporal Logic to reason about these games.
• Cryptanalysis of Symmetric Primitives over Rings and a Key Recovery Attack on Rubato
  
  • Grassi Lorenzo
  • Manterola Ayala Irati
  • Hovd Martha Norberg
  • Øygarden Morten
  • Raddum Håvard
  • Wang Qingju
  , 2023, 14083, pp.305-339. Symmetric primitives are a cornerstone of cryptography, and have traditionally been defined over fields, where cryptanalysis is now well understood. However, a few symmetric primitives defined over rings Zq for a composite number q have recently been proposed, a setting where security is much less studied. In this paper we focus on studying established algebraic attacks typically defined over fields and the extent of their applicability to symmetric primitives defined over the ring of integers modulo a composite q Based on our analysis, we present an attack on full Rubato, a family of symmetric ciphers proposed by Ha et al. at Eurocrypt 2022 designed to be used in a transciphering framework for approximate fully homomorphic encryption. We show that at least 25 % of the possible choices for q satisfy certain conditions that lead to a successful key recovery attack with complexity significantly lower than the claimed security level for five of the six ciphers in the Rubato family. (10.1007/978-3-031-38548-3_11)
  DOI : 10.1007/978-3-031-38548-3_11
• Horst Meets Fluid-SPN: Griffin for Zero-Knowledge Applications
  
  • Grassi Lorenzo
  • Hao Yonglin
  • Rechberger Christian
  • Schofnegger Markus
  • Walch Roman
  • Wang Qingju
  , 2023, 14083, pp.573-606. Zero-knowledge (ZK) applications form a large group of use cases in modern cryptography, and recently gained in popularity due to novel proof systems. For many of these applications, cryptographic hash functions are used as the main building blocks, and they often dominate the overall performance and cost of these approaches. Therefore, in the last years several new hash functions were built in order to reduce the cost in these scenarios, including POSEIDON and Rescue among others. These hash functions often look very different from more classical designs such as AES or SHA-2. For example, they work natively over prime fields rather than binary ones. At the same time, for example POSEIDON and Rescue share some common features, such as being SPN schemes and instantiating the nonlinear layer with invertible power maps. While this allows the designers to provide simple and strong arguments for establishing their security, it also introduces crucial limitations in the design, which may affect the performance in the target applications. In this paper, we propose the Horst construction, in which the addition in a Feistel scheme (x,y)\mapsto (y+F(x), x) is extended via a multiplication, i.e., (x,y)\mapsto (y\times G(x) + F(x), x). By carefully analyzing the performance metrics in SNARK and STARK protocols, we show how to combine an expanding Horst scheme with a Rescue-like SPN scheme in order to provide security and better efficiency in the target applications. We provide an extensive security analysis for our new design GRIFFIN and a comparison with all current competitors. (10.1007/978-3-031-38548-3_19)
  DOI : 10.1007/978-3-031-38548-3_19
• Automatic Data Augmentation for Domain Adapted Fine-Tuning of Self-Supervised Speech Representations
  
  • Zaiem Salah
  • Parcollet Titouan
  • Essid Slim
  , 2023, pp.67-71. Self-Supervised Learning (SSL) has allowed leveraging large amounts of unlabeled speech data to improve the performance of speech recognition models even with small annotated datasets. Despite this, speech SSL representations may fail while facing an acoustic mismatch between the pretraining and target datasets. To address this issue, we propose a novel supervised domain adaptation method, designed for cases exhibiting such a mismatch in acoustic domains. It consists in applying properly calibrated data augmentations on a large clean dataset, bringing it closer to the target domain, and using it as part of an initial fine-tuning stage. Augmentations are automatically selected through the minimization of a conditional-dependence estimator, based on the target dataset. The approach is validated during an oracle experiment with controlled distortions and on two amateur-collected low-resource domains, reaching better performances compared to the baselines in both cases. (10.21437/Interspeech.2023-1040)
  DOI : 10.21437/Interspeech.2023-1040
• Speech Self-Supervised Representation Benchmarking: Are We Doing it Right?
  
  • Zaiem Salah
  • Kemiche Youcef
  • Parcollet Titouan
  • Essid Slim
  • Ravanelli Mirco
  , 2023, pp.2873-2877. Self-supervised learning (SSL) has recently allowed leveraging large datasets of unlabeled speech signals to reach impressive performance on speech tasks using only small amounts of annotated data. The high number of proposed approaches fostered the need and rise of extended benchmarks that evaluate their performance on a set of downstream tasks exploring various aspects of the speech signal. However, and while the number of considered tasks has been growing, most rely upon a single decoding architecture that maps the frozen SSL representations to the downstream labels. This work investigates the robustness of such benchmarking results to changes in the decoder architecture. Interestingly, it appears that varying the architecture of the downstream decoder leads to significant variations in the leaderboards of most tasks. Concerningly, our study reveals that benchmarking using limited decoders may cause a counterproductive increase in the sizes of the developed SSL models. (10.21437/Interspeech.2023-1087)
  DOI : 10.21437/Interspeech.2023-1087
• ESWC 2023 Workshops and Tutorials Joint Proceedings
  
  • Alam Mehwish
  • Trojahn Cassia
  • Hertling Sven
  • Pesquita Catia
  • Aebeloe Christian
  • Aras Hidir
  • Azzam Amr
  • Cano Juan
  • Domingue John
  • Gottschalk Simon
  • Hartig Olaf
  • Hose Katja
  • Kirrane Sabrina
  • Lisena Pasquale
  • Osborne Francesco
  • Rohde Philipp
  • Steels Luc
  • Taelman Ruben
  • Third Aisling
  • Tiddi Ilaria
  • Türker Rima
  , 2023, 3443. Joint Proceedings of the ESWC 2023 Workshops and Tutorials co-located with the 20th European Semantic Web Conference (ESWC 2023)
• Survey on Online Streaming Continual Learning
  
  • Gunasekara Nuwan
  • Pfahringer Bernhard
  • Gomes Heitor Murilo
  • Bifet Albert
  , 2023, pp.6628--6637. Stream Learning (SL) attempts to learn from a data stream efficiently. A data stream learning algorithm should adapt to input data distribution shifts without sacrificing accuracy. These distribution shifts are known as ”concept drifts” in the literature. SL provides many supervised, semi-supervised, and unsupervised methods for detecting and adjusting to concept drift. On the other hand, Continual Learning (CL) attempts to preserve previous knowledge while performing well on the current concept when confronted with concept drift. In Online Continual Learning (OCL), this learning happens online. This survey explores the intersection of those two online learning paradigms to find synergies. We identify this intersection as Online Streaming Continual Learning (OSCL). The study starts with a gentle introduction to SL and then explores CL. Next, it explores OSCL from SL and OCL perspectives to point out new research trends and give directions for future research. (10.24963/IJCAI.2023/743)
  DOI : 10.24963/IJCAI.2023/743
• RF Electromagnetic Fields Exposure Monitoring using Drive Test and Sensors in a French City
  
  • Wang Shanshan
  • Chikha Wassim Ben
  • Zhang Yarui
  • Liu Jiang
  • Conil Emmanuelle
  • Jawad Ourouk
  • Ourak Lamine
  • Wiart Joe
  , 2023, pp.1-4. This paper presents the monitoring work on radiofrequency (RF) electromagnetic field (EMF) exposure, carried out in a French city by using drive test and sensor network methods. The drive test measurements are realized by continuously measuring in a car via a portable spectrum analyzer, i.e., Tektronix RSA 306B, connected to a 3-axis antenna. It records electric (E) field values over a broad cellular band (700 MHz – 3.8 GHz) along the pre-defined outdoor routes at a height of 2 meters above the soil level. Secondly, a total of 19 sensors are installed near the center of the city, on the streetlamps (4 meters to the soil), which measure the broad band (250 kHz – 6 GHz) exposure level. The statistical distribution of measurement data is analyzed, and compared with base station antenna (BSA) information, which can be publicly accessed. Furthermore, the temporal variation of measurements collected by sensor networks is analyzed. (10.23919/URSIGASS57860.2023.10265402)
  DOI : 10.23919/URSIGASS57860.2023.10265402
• Impact of sampling frequency on the performance of DEVIN: A personal EM UL exposimeter
  
  • Mazloum Taghrid
  • Bories Serge
  • Dassonville David
  • Wiart Joe
  , 2023, pp.1-4. DEVIN is a miniaturized personal electromagnetic (EM) exposimeter that allows recording powers emitted by the mobile phone as well as identifying the user activities along the day. It is simply a hardware integrated into the shell of the smartphone. In the context of epidemiological studies, DEVIN is required to collect data from volunteers over a long period of time, e.g., a week. This implies a critical challenge related to the autonomy of its battery, which should be charged once per day. One solution to extend the battery autonomy consists on decreasing the sampling frequency. Therefore we intend in the present work to evaluate the performance of DEVIN at two different sampling frequencies for two various signals types, i.e., voice over LTE (VoLTE) and data uploading signals. (10.23919/URSIGASS57860.2023.10265631)
  DOI : 10.23919/URSIGASS57860.2023.10265631
• Anamorphic Signatures: Secrecy from a Dictator Who Only Permits Authentication!
  
  • Kutyłowski Mirosław
  • Persiano Giuseppe
  • Phan Duong Hieu
  • Yung Moti
  • Zawada Marcin
  , 2023, 14082, pp.759-790. The goal of this research is to raise technical doubts regarding the usefulness of the repeated attempts by governments to curb Cryptography (aka the "Crypto Wars"), and argue that they, in fact, cause more damage than adding effective control. The notion of Anamorphic Encryption was presented in Eurocrypt'22 for a similar aim. There, despite the presence of a Dictator who possesses all keys and knows all messages, parties can arrange a hidden "anamorphic" message in an otherwise indistinguishable from regular ciphertexts (wrt the Dictator). In this work, we postulate a stronger cryptographic control setting where encryption does not exist (or is neutralized) since all communication is passed through the Dictator in, essentially, cleartext mode (or otherwise, when secure channels to and from the Dictator are the only confidentiality mechanism). Messages are only authenticated to assure recipients of the identity of the sender. We ask whether security against the Dictator still exists, even under such a strict regime which allows only authentication (i.e., authenticated/ signed messages) to pass end-toend, and where received messages are determined by/ known to the Dictator, and the Dictator also eventually gets all keys to verify compliance of past signing. To frustrate the Dictator, this authenticated message setting gives rise to the possible notion of anamorphic channels inside signature and authentication schemes, where parties attempt to send undetectable secure messages (or other values) using signature tags which are indistinguishable from regular tags. We define and present implementation of schemes for anamorphic signature and authentication; these are applicable to existing and standardized signature and authentication schemes which were designed independently of the notion of anamorphic messages. Further, some cornerstone constructions of the foundations of signatures, in fact, introduce anamorphism. (10.1007/978-3-031-38545-2_25)
  DOI : 10.1007/978-3-031-38545-2_25
• Scalable Verification of Strategy Logic through Three-Valued Abstraction
  
  • Belardinelli Francesco
  • Ferrando Angelo
  • Jamroga Wojciech
  • Malvone Vadim
  • Murano Aniello
  , 2023, pp.46-54. The model checking problem for multi-agent systems against Strategy Logic specifications is known to be non-elementary. On this logic several fragments have been defined to tackle this issue but at the expense of expressiveness. In this paper, we propose a three-valued semantics for Strategy Logic upon which we define an abstraction method. We show that the latter semantics is an approximation of the classic two-valued one for Strategy Logic. Furthermore, we extend MCMAS, an open-source model checker for multi-agent specifications, to incorporate our abstraction method and present some promising experimental results. (10.24963/IJCAI.2023/6)
  DOI : 10.24963/IJCAI.2023/6
• OpenAI Cribbed Our Tax Example, But Can GPT-4 Really Do Tax?
  
  • Blair-Stanek Andrew
  • Holzenberger Nils
  • van Durme Benjamin
  Tax Notes Federal, Tax Analysts, 2023. The authors explain where OpenAI got the tax law example in its livestream demonstration of GPT-4, why GPT-4 got the wrong answer, and how it fails to reliably calculate taxes. (10.48550/arXiv.2309.09992)
  DOI : 10.48550/arXiv.2309.09992
• On the Fair Comparison of Optimization Algorithms in Different Machines
  
  • Arza Etor
  • Ceberio Josu
  • Irurozki Ekhine
  • Perez Aritz
  Annals of Applied Statistics, Institute of Mathematical Statistics, 2023. An experimental comparison of two or more optimization algorithms requires the same computational resources to be assigned to each algorithm. When a maximum runtime is set as the stopping criterion, all algorithms need to be executed in the same machine if they are to use the same resources. Unfortunately, the implementation code of the algorithms is not always available, which means that running the algorithms to be compared in the same machine is not always possible. And even if they are available, some optimization algorithms might be costly to run, such as training large neural-networks in the cloud. In this paper, we consider the following problem: how do we compare the performance of a new optimization algorithm B with a known algorithm A in the literature if we only have the results (the objective values) and the runtime in each instance of algorithm A? Particularly, we present a methodology that enables a statistical analysis of the performance of algorithms executed in different machines. The proposed methodology has two parts. First, we propose a model that, given the runtime of an algorithm in a machine, estimates the runtime of the same algorithm in another machine. This model can be adjusted so that the probability of estimating a runtime longer than what it should be is arbitrarily low. Second, we introduce an adaptation of the one-sided sign test that uses a modified p-value and takes into account that probability. Such adaptation avoids increasing the probability of type I error associated with executing algorithms A and B in different machines.
• Who's speaking? Predicting speaker profession from speech
  
  • Wu Yaru
  • Chen Lihu
  • Elie Benjamin
  • Suchanek Fabian M.
  • Vasilescu Ioana
  • Lamel Lori
  , 2023, pp.3086-3090. Variations in speech can reveal the gender, birth place, age, and socioeconomic level of the speaker. In this paper, we show that even the profession of the speaker can be recovered from a recording. For this purpose, we design a method that combines features from both the speech signal and the transcription. For the features from the transcription, we used pretrained language models. This allows us to train a model that predicts the speaker profession from both signals. Our empirical results show that our model can narrow down the profession of the speakers considerably.
• MesoGen: Designing Procedural On-Surface Stranded Mesostructures
  
  • Michel Élie
  • Boubekeur Tamy
  , 2023 (50), pp.1-10. Three-dimensional mesostructures enrich coarse macrosurfaces with complex features, which are 3D geometry with arbitrary topology in essence, but are expected to be self-similar with no tiling artifacts, just like texture-based material models. This is a challenging task, as no existing modeling tool provides the right constraints in the design phase to ensure such properties while maintaining real-time editing capabilities. In this paper, we propose MesoGen, a novel tile-centric authoring approach for the design of procedural mesostructures featuring non-periodic self-similarity while being represented as a compact and GPU-friendly model. We ensure by construction the continuity of the mesostructure: the user designs a set of atomic tiles by drawing 2D cross-sections on the interfaces between tiles, and selecting pairs of cross-sections to be connected as strands, i.e., 3D sweep surfaces. In parallel, a tiling engine continuously fills the shell space of the macrosurface with the so-defined tile set while ensuring that only matching interfaces are in contact. Moreover, the engine suggests to the user the addition of new tiles whenever the problem happens to be over-constrained. As a result, our method allows for the rapid creation of complex, seamless procedural mesostructure and is particularly adapted for wicker-like ones, often impossible to achieve with scattering-based mesostructure synthesis methods. (10.1145/3588432.3591496)
  DOI : 10.1145/3588432.3591496
• Variational Shape Reconstruction via Quadric Error Metrics
  
  • Zhao Tong
  • Busé Laurent
  • Cohen-Steiner David
  • Boubekeur Tamy
  • Thiery Jean-Marc
  • Alliez Pierre
  , 2023. Inspired by the strengths of quadric error metrics initially designed for mesh decimation, we propose a concise mesh reconstruction approach for 3D point clouds. Our approach proceeds by clustering the input points enriched with quadric error metrics, where the generator of each cluster is the optimal 3D point for the sum of its quadric error metrics. This approach favors the placement of generators on sharp features, and tends to equidistribute the error among clusters. We reconstruct the output surface mesh from the adjacency between clusters and a constrained binary solver. We combine our clustering process with an adaptive refinement driven by the error. Compared to prior art, our method avoids dense reconstruction prior to simplification and produces immediately an optimized mesh. (10.1145/3588432.3591529)
  DOI : 10.1145/3588432.3591529
• W-Sec: a Model-Based Formal Method for Assessing the Impacts of Security Countermeasures
  
  • Sultan Bastien
  • Apvrille Ludovic
  • Jaillon Philippe
  • Coudert Sophie
  , 2023, 1708, pp.203-229. The chapter provides a detailed description of W-Sec, a formal model-based countermeasures' impact assessment method. It also introduces a new formal definition of the two SysML profiles used in SysML-Sec and W-Sec, enabling (i) for the future automation of several W-Sec stages and (ii) for the definition of consistency rules ensuring the consistency of the models written in these two distinct modeling languages. In addition, the chapter evaluates W-Sec with a new industry 4.0 case-study and discusses the strengths and the current limitations of the approach in this new application field. (10.1007/978-3-031-38821-7_10)
  DOI : 10.1007/978-3-031-38821-7_10
• Dependency Graphs to Boost the Verification of SysML Models
  
  • Apvrille Ludovic
  • de Saqui-Sannes Pierre
  • Hotescu Oana
  • Calvino Alessandro Tempia
  Communications in Computer and Information Science, Springer Verlag, 2023, 1708, pp.109-134. Model-Based Systems Engineering has often been associated with the Systems Modeling Language. Several SysML tools offer formal verification ca- pabilities, and therefore enable early detection of design errors in the life cycle of systems. Model-checking is a common formal verification approach used to assess the satisfiability of properties. Thus, a SysML model and a property can be injected into a model-checker returning a true/false result. A drawback of this approach is that the entire SysML model is used for the verification, even if the property targets a sub-system of the model. In this paper, it is suggested to rely on dependency graphs to avoid applying model checking to the entire system when only a subset of the latter needs to be taken into account. We formalize SysML models and properties, then we present new algorithms to generate and reduce de- pendency graphs, so as to perform verification on reduced models. A case study on Time-Sensitive Networking is used to demonstrate the efficiency and limits of this approach. The algorithms described in the paper are fully implemented by the free software TTool. Our method enables an improvement in run time between 3% and 90% depending on the state space to be traversed to verify the property. (10.1007/978-3-031-38821-7_6)
  DOI : 10.1007/978-3-031-38821-7_6
• PyClause -Simple and Efficient Rule Handling for Knowledge Graphs
  
  • Betz Patrick
  • Galárraga Luis
  • Ott Simon
  • Meilicke Christian
  • Suchanek Fabian
  • Stuckenschmidt Heiner
  , 2024, pp.8610-8613. <div><p>Rule mining finds patterns in structured data such as knowledge graphs. Rules can predict facts, help correct errors, and yield explainable insights about the data. However, existing rule mining implementations focus exclusively on mining rules -and not on their application. The PyClause library offers a rich toolkit for the application of the mined rules: from explaining facts to predicting links, scoring rules, and deducing query results. The library is easy to use and can handle substantial data loads.</p></div> (10.24963/ijcai.2024/991)
  DOI : 10.24963/ijcai.2024/991
• Materials Fatigue Prediction Using Graph Neural Networks on Microstructure Representations
  
  • Thomas Akhil
  • Durmaz Ali Riza
  • Alam Mehwish
  • Gumbsch Peter
  • Sack Harald
  • Eberl Chris
  Scientific Reports, Nature Publishing Group, 2023, 13 (1), pp.12562. The local prediction of fatigue damage within polycrystals in a high-cycle fatigue setting is a long-lasting and challenging task. It requires identifying grains tending to accumulate plastic deformation under cyclic loading. We address this task by transcribing ferritic steel microtexture and damage maps from experiments into a microstructure graph. Here, grains constitute graph nodes connected by edges whenever grains share a common boundary. Fatigue loading causes some grains to develop slip markings, which can evolve into microcracks and lead to failure. This data set enables applying graph neural network variants on the task of binary grain-wise damage classification. The objective is to identify suitable data representations and models with an appropriate inductive bias to learn the underlying damage formation causes. Here, graph convolutional networks yielded the best performance with a balanced accuracy of 0.72 and an F1-score of 0.34, outperforming phenomenological crystal plasticity (+ 68%) and conventional machine learning (+ 17%) models by large margins. Further, we present an interpretability analysis that highlights the grains along with features that are considered important by the graph model for the prediction of fatigue damage initiation, thus demonstrating the potential of such techniques to reveal underlying mechanisms and microstructural driving forces in critical grain ensembles. (10.1038/s41598-023-39400-2)
  DOI : 10.1038/s41598-023-39400-2
• Testing and reliability enhancement of security primitives: Methodology and experimental validation
  
  • Anik Md Toufiq Hasan
  • Danger Jean-Luc
  • Diankha Omar
  • Ebrahimabadi Mohammad
  • Frisch Christoph
  • Guilley Sylvain
  • Karimi Naghmeh
  • Pehl Michael
  • Takarabt Sofiane
  Microelectronics Reliability, Elsevier, 2023, 147, pp.115055. The test of security primitives is particularly strategic as any bias coming from the implementation or environment can wreak havoc on the security it is intended to provide. This paper presents how some security properties are tested on hardware security primitives including True Random Number Generation (TRNG), Physically Unclonable Function (PUF), and cryptographic modules. Moreover, we discuss how the sensors embedded to protect cryptographic modules against fault injection attacks should be calibrated over time to fulfill the requirement it was designed for. The testing we discuss in this paper is different from the conventional testing where we consider a fault model and generate test patterns via an ATPG to detect such faults. The test of TRNG and PUF to ensure a high level of security is mainly about the entropy assessment, which requires specific statistical tests. The security against side-channel analysis (SCA) of cryptographic primitives, like the substitution box in symmetric cryptography, is generally ensured by masking. However, the hardware implementation of masking can be damaged by glitches, which create leakages on sensitive variables. Accordingly, a test method is to search for nets of the cryptographic netlist, which are vulnerable to glitches. Finally, the Digital Sensor (DS) is an efficient primitive to detect disturbances and raise alarms in the case of fault injection attack (FIA). The dimensioning of this primitive requires a precise test to take into account the environmental variations including aging. This paper extends on a conference paper presented at DFTS’21 by the same co-authors, where the test methodology for three critical security primitives is presented. In addition, in this paper, we add experimental validation to show how such testing methodology is applied in practice. (10.1016/j.microrel.2023.115055)
  DOI : 10.1016/j.microrel.2023.115055