arXiv subjects

Stjepan Picek

Publications and source records attributed to Stjepan Picek.

At least 91 records · Page 5Linked to original sources

arXiv2022

Watermarking Graph Neural Networks based on Backdoor Attacks

Graph Neural Networks (GNNs) have achieved promising performance in various real-world applications. Building a powerful GNN model is not a trivial task, as it requires a large amount of training data, powerful computing resources, and human expertise in fine-tuning the model. Moreover, with the development of adversarial attacks, e.g., model stealing attacks, GNNs raise challenges to model authentication. To avoid copyright infringement on GNNs, verifying the ownership of the GNN models is necessary. This paper presents a watermarking framework for GNNs for both graph and node classification tasks. We 1) design two strategies to generate watermarked data for the graph classification task and one for the node classification task, 2) embed the watermark into the host model through training to obtain the watermarked GNN model, and 3) verify the ownership of the suspicious model in a black-box setting. The experiments show that our framework can verify the ownership of GNN models with a very high probability (up to $99\%$) for both tasks. Finally, we experimentally show that our watermarking approach is robust against a state-of-the-art model extraction technique and four state-of-the-art defenses against backdoor attacks.

cs.LG↗

arXiv2022

On the Vulnerability of Data Points under Multiple Membership Inference Attacks and Target Models

Membership Inference Attacks (MIAs) infer whether a data point is in the training data of a machine learning model. It is a threat while being in the training data is private information of a data point. MIA correctly infers some data points as members or non-members of the training data. Intuitively, data points that MIA accurately detects are vulnerable. Considering those data points may exist in different target models susceptible to multiple MIAs, the vulnerability of data points under multiple MIAs and target models is worth exploring. This paper defines new metrics that can reflect the actual situation of data points' vulnerability and capture vulnerable data points under multiple MIAs and target models. From the analysis, MIA has an inference tendency to some data points despite a low overall inference performance. Additionally, we implement 54 MIAs, whose average attack accuracy ranges from 0.5 to 0.9, to support our analysis with our scalable and flexible platform, Membership Inference Attacks Platform (VMIAP). Furthermore, previous methods are unsuitable for finding vulnerable data points under multiple MIAs and different target models. Finally, we observe that the vulnerability is not characteristic of the data point but related to the MIA and target model.

cs.CR↗

arXiv2022

DARWIN: Survival of the Fittest Fuzzing Mutators

Fuzzing is an automated software testing technique broadly adopted by the industry. A popular variant is mutation-based fuzzing, which discovers a large number of bugs in practice. While the research community has studied mutation-based fuzzing for years now, the algorithms' interactions within the fuzzer are highly complex and can, together with the randomness in every instance of a fuzzer, lead to unpredictable effects. Most efforts to improve this fragile interaction focused on optimizing seed scheduling. However, real-world results like Google's FuzzBench highlight that these approaches do not consistently show improvements in practice. Another approach to improve the fuzzing process algorithmically is optimizing mutation scheduling. Unfortunately, existing mutation scheduling approaches also failed to convince because of missing real-world improvements or too many user-controlled parameters whose configuration requires expert knowledge about the target program. This leaves the challenging problem of cleverly processing test cases and achieving a measurable improvement unsolved. We present DARWIN, a novel mutation scheduler and the first to show fuzzing improvements in a realistic scenario without the need to introduce additional user-configurable parameters, opening this approach to the broad fuzzing community. DARWIN uses an Evolution Strategy to systematically optimize and adapt the probability distribution of the mutation operators during fuzzing. We implemented a prototype based on the popular general-purpose fuzzer AFL. DARWIN significantly outperforms the state-of-the-art mutation scheduler and the AFL baseline in our own coverage experiment, in FuzzBench, and by finding 15 out of 21 bugs the fastest in the MAGMA benchmark. Finally, DARWIN found 20 unique bugs (including one novel bug), 66% more than AFL, in widely-used real-world applications.

cs.SE↗

arXiv2022

Label-Only Membership Inference Attack against Node-Level Graph Neural Networks

Graph Neural Networks (GNNs), inspired by Convolutional Neural Networks (CNNs), aggregate the message of nodes' neighbors and structure information to acquire expressive representations of nodes for node classification, graph classification, and link prediction. Previous studies have indicated that GNNs are vulnerable to Membership Inference Attacks (MIAs), which infer whether a node is in the training data of GNNs and leak the node's private information, like the patient's disease history. The implementation of previous MIAs takes advantage of the models' probability output, which is infeasible if GNNs only provide the prediction label (label-only) for the input. In this paper, we propose a label-only MIA against GNNs for node classification with the help of GNNs' flexible prediction mechanism, e.g., obtaining the prediction label of one node even when neighbors' information is unavailable. Our attacking method achieves around 60\% accuracy, precision, and Area Under the Curve (AUC) for most datasets and GNN models, some of which are competitive or even better than state-of-the-art probability-based MIAs implemented under our environment and settings. Additionally, we analyze the influence of the sampling method, model selection approach, and overfitting level on the attack performance of our label-only MIA. Both of those factors have an impact on the attack performance. Then, we consider scenarios where assumptions about the adversary's additional dataset (shadow dataset) and extra information about the target model are relaxed. Even in those scenarios, our label-only MIA achieves a better attack performance in most cases. Finally, we explore the effectiveness of possible defenses, including Dropout, Regularization, Normalization, and Jumping knowledge. None of those four defenses prevent our attack completely.

cs.CR↗

arXiv2022

On the Security & Privacy in Federated Learning

Recent privacy awareness initiatives such as the EU General Data Protection Regulation subdued Machine Learning (ML) to privacy and security assessments. Federated Learning (FL) grants a privacy-driven, decentralized training scheme that improves ML models' security. The industry's fast-growing adaptation and security evaluations of FL technology exposed various vulnerabilities that threaten FL's confidentiality, integrity, or availability (CIA). This work assesses the CIA of FL by reviewing the state-of-the-art (SoTA) and creating a threat model that embraces the attack's surface, adversarial actors, capabilities, and goals. We propose the first unifying taxonomy for attacks and defenses and provide promising future research directions.

cs.CR↗

arXiv2022

Can You Hear It? Backdoor Attacks via Ultrasonic Triggers

This work explores backdoor attacks for automatic speech recognition systems where we inject inaudible triggers. By doing so, we make the backdoor attack challenging to detect for legitimate users, and thus, potentially more dangerous. We conduct experiments on two versions of a speech dataset and three neural networks and explore the performance of our attack concerning the duration, position, and type of the trigger. Our results indicate that less than 1% of poisoned data is sufficient to deploy a backdoor attack and reach a 100% attack success rate. We observed that short, non-continuous triggers result in highly successful attacks. However, since our trigger is inaudible, it can be as long as possible without raising any suspicions making the attack more effective. Finally, we conducted our attack in actual hardware and saw that an adversary could manipulate inference in an Android application by playing the inaudible trigger over the air.

cs.CR↗

arXiv2022

Dynamic Backdoors with Global Average Pooling

Outsourced training and machine learning as a service have resulted in novel attack vectors like backdoor attacks. Such attacks embed a secret functionality in a neural network activated when the trigger is added to its input. In most works in the literature, the trigger is static, both in terms of location and pattern. The effectiveness of various detection mechanisms depends on this property. It was recently shown that countermeasures in image classification, like Neural Cleanse and ABS, could be bypassed with dynamic triggers that are effective regardless of their pattern and location. Still, such backdoors are demanding as they require a large percentage of poisoned training data. In this work, we are the first to show that dynamic backdoor attacks could happen due to a global average pooling layer without increasing the percentage of the poisoned training data. Nevertheless, our experiments in sound classification, text sentiment analysis, and image classification show this to be very difficult in practice.

cs.CR↗

arXiv2022

Evolving Constructions for Balanced, Highly Nonlinear Boolean Functions

Finding balanced, highly nonlinear Boolean functions is a difficult problem where it is not known what nonlinearity values are possible to be reached in general. At the same time, evolutionary computation is successfully used to evolve specific Boolean function instances, but the approach cannot easily scale for larger Boolean function sizes. Indeed, while evolving smaller Boolean functions is almost trivial, larger sizes become increasingly difficult, and evolutionary algorithms perform suboptimally. In this work, we ask whether genetic programming (GP) can evolve constructions resulting in balanced Boolean functions with high nonlinearity. This question is especially interesting as there are only a few known such constructions. Our results show that GP can find constructions that generalize well, i.e., result in the required functions for multiple tested sizes. Further, we show that GP evolves many equivalent constructions under different syntactic representations. Interestingly, the simplest solution found by GP is a particular case of the well-known indirect sum construction.

cs.NE↗

arXiv2022

Modeling Strong Physically Unclonable Functions with Metaheuristics

Evolutionary algorithms have been successfully applied to attacking Physically Unclonable Functions (PUFs). CMA-ES is recognized as the most powerful option for a type of attack called the reliability attack. While there is no reason to doubt the performance of CMA-ES, the lack of comparison with different metaheuristics and results for the challenge-response pair-based attack leaves open questions if there are better-suited metaheuristics for the problem. In this paper, we take a step back and systematically evaluate several metaheuristics for the challenge-response pair-based attack on strong PUFs. Our results confirm that CMA-ES has the best performance, but we also note several other algorithms with similar performance while having smaller computational costs. More precisely, if we provide a sufficient number of challenge-response pairs to train the algorithm, various configurations show good results. Consequently, we conclude that EAs represent a strong option for challenge-response pair-based attacks on PUFs.

cs.NE↗

arXiv2022

Evolutionary Construction of Perfectly Balanced Boolean Functions

Finding Boolean functions suitable for cryptographic primitives is a complex combinatorial optimization problem, since they must satisfy several properties to resist cryptanalytic attacks, and the space is very large, which grows super exponentially with the number of input variables. Recent research has focused on the study of Boolean functions that satisfy properties on restricted sets of inputs due to their importance in the development of the FLIP stream cipher. In this paper, we consider one such property, perfect balancedness, and investigate the use of Genetic Programming (GP) and Genetic Algorithms (GA) to construct Boolean functions that satisfy this property along with a good nonlinearity profile. We formulate the related optimization problem and define two encodings for the candidate solutions, namely the truth table and the weightwise balanced representations. Somewhat surprisingly, the results show that GA with the weightwise balanced representation outperforms GP with the classical truth table phenotype in finding highly nonlinear WPB functions. This finding is in stark contrast to previous findings on the evolution of globally balanced Boolean functions, where GP always performs best.

cs.NE↗

arXiv2021

Bent Functions in the Partial Spread Class Generated by Linear Recurring Sequences

We present a construction of partial spread bent functions using subspaces generated by linear recurring sequences (LRS). We first show that the kernels of the linear mappings defined by two LRS have a trivial intersection if and only if their feedback polynomials are relatively prime. Then, we characterize the appropriate parameters for a family of pairwise coprime polynomials to generate a partial spread required for the support of a bent function, showing that such families exist if and only if the degrees of the underlying polynomials is either $1$ or $2$. We then count the resulting sets of polynomials and prove that for degree $1$, our LRS construction coincides with the Desarguesian partial spread. Finally, we perform a computer search of all $\mathcal{PS}^-$ and $\mathcal{PS}^+$ bent functions of $n=8$ variables generated by our construction and compute their 2-ranks. The results show that many of these functions defined by polynomials of degree $b=2$ are not EA-equivalent to any Maiorana-McFarland or Desarguesian partial spread function.

cs.CR↗

arXiv2021

Hand Me Your PIN! Inferring ATM PINs of Users Typing with a Covered Hand

Automated Teller Machines (ATMs) represent the most used system for withdrawing cash. The European Central Bank reported more than 11 billion cash withdrawals and loading/unloading transactions on the European ATMs in 2019. Although ATMs have undergone various technological evolutions, Personal Identification Numbers (PINs) are still the most common authentication method for these devices. Unfortunately, the PIN mechanism is vulnerable to shoulder-surfing attacks performed via hidden cameras installed near the ATM to catch the PIN pad. To overcome this problem, people get used to covering the typing hand with the other hand. While such users probably believe this behavior is safe enough to protect against mentioned attacks, there is no clear assessment of this countermeasure in the scientific literature. This paper proposes a novel attack to reconstruct PINs entered by victims covering the typing hand with the other hand. We consider the setting where the attacker can access an ATM PIN pad of the same brand/model as the target one. Afterward, the attacker uses that model to infer the digits pressed by the victim while entering the PIN. Our attack owes its success to a carefully selected deep learning architecture that can infer the PIN from the typing hand position and movements. We run a detailed experimental analysis including 58 users. With our approach, we can guess 30% of the 5-digit PINs within three attempts -- the ones usually allowed by ATM before blocking the card. We also conducted a survey with 78 users that managed to reach an accuracy of only 7.92% on average for the same setting. Finally, we evaluate a shielding countermeasure that proved to be rather inefficient unless the whole keypad is shielded.

cs.CR↗

arXiv2021

On the Difficulty of Evolving Permutation Codes

Combinatorial designs provide an interesting source of optimization problems. Among them, permutation codes are particularly interesting given their applications in powerline communications, flash memories, and block ciphers. This paper addresses the design of permutation codes by evolutionary algorithms (EA) by developing an iterative approach. Starting from a single random permutation, new permutations satisfying the minimum distance constraint are incrementally added to the code by using a permutation-based EA. We investigate our approach against four different fitness functions targeting the minimum distance requirement at different levels of detail and with two different policies concerning code expansion and pruning. We compare the results achieved by our EA approach to those of a simple random search, remarking that neither method scales well with the problem size.

cs.NE↗

arXiv2021

Explainability-based Backdoor Attacks Against Graph Neural Networks

Backdoor attacks represent a serious threat to neural network models. A backdoored model will misclassify the trigger-embedded inputs into an attacker-chosen target label while performing normally on other benign inputs. There are already numerous works on backdoor attacks on neural networks, but only a few works consider graph neural networks (GNNs). As such, there is no intensive research on explaining the impact of trigger injecting position on the performance of backdoor attacks on GNNs. To bridge this gap, we conduct an experimental investigation on the performance of backdoor attacks on GNNs. We apply two powerful GNN explainability approaches to select the optimal trigger injecting position to achieve two attacker objectives -- high attack success rate and low clean accuracy drop. Our empirical results on benchmark datasets and state-of-the-art neural network models demonstrate the proposed method's effectiveness in selecting trigger injecting position for backdoor attacks on GNNs. For instance, on the node classification task, the backdoor attack with trigger injecting position selected by GraphLIME reaches over $84 \%$ attack success rate with less than $2.5 \%$ accuracy drop

cs.LG↗

arXiv2021

Evolutionary Algorithms for Designing Reversible Cellular Automata

Reversible Cellular Automata (RCA) are a particular kind of shift-invariant transformations characterized by a dynamics composed only of disjoint cycles. They have many applications in the simulation of physical systems, cryptography and reversible computing. In this work, we formulate the search of a specific class of RCA -- namely, those whose local update rules are defined by conserved landscapes -- as an optimization problem to be tackled with Genetic Algorithms (GA) and Genetic Programming (GP). In particular, our experimental investigation revolves around three different research questions, which we address through a single-objective, a multi-objective, and a lexicographic approach. The results obtained from our experiments corroborate the previous findings and shed new light on 1) the difficulty of the associated optimization problem for GA and GP, 2) the relevance of conserved landscape CA in the domain of cryptography and reversible computing, and 3) the relationship between the reversibility property and the Hamming weight.

cs.NE↗

arXiv2021

On the Genotype Compression and Expansion for Evolutionary Algorithms in the Continuous Domain

This paper investigates the influence of genotype size on evolutionary algorithms' performance. We consider genotype compression (where genotype is smaller than phenotype) and expansion (genotype is larger than phenotype) and define different strategies to reconstruct the original variables of the phenotype from both the compressed and expanded genotypes. We test our approach with several evolutionary algorithms over three sets of optimization problems: COCO benchmark functions, modeling of Physical Unclonable Functions, and neural network weight optimization. Our results show that genotype expansion works significantly better than compression, and in many scenarios, outperforms the original genotype encoding. This could be attributed to the change in the genotype-phenotype mapping introduced with the expansion methods: this modification beneficially transforms the domain landscape and alleviates the search space traversal.

cs.NE↗

arXiv2021

CoInGP: Convolutional Inpainting with Genetic Programming

We investigate the use of Genetic Programming (GP) as a convolutional predictor for missing pixels in images. The training phase is performed by sweeping a sliding window over an image, where the pixels on the border represent the inputs of a GP tree. The output of the tree is taken as the predicted value for the central pixel. We consider two topologies for the sliding window, namely the Moore and the Von Neumann neighborhood. The best GP tree scoring the lowest prediction error over the training set is then used to predict the pixels in the test set. We experimentally assess our approach through two experiments. In the first one, we train a GP tree over a subset of 1000 complete images from the MNIST dataset. The results show that GP can learn the distribution of the pixels with respect to a simple baseline predictor, with no significant differences observed between the two neighborhoods. In the second experiment, we train a GP convolutional predictor on two degraded images, removing around 20% of their pixels. In this case, we observe that the Moore neighborhood works better, although the Von Neumann neighborhood allows for a larger training set.

cs.NE↗

arXiv2020

ReproducedPapers.org: Openly teaching and structuring machine learning reproducibility

We present ReproducedPapers.org: an open online repository for teaching and structuring machine learning reproducibility. We evaluate doing a reproduction project among students and the added value of an online reproduction repository among AI researchers. We use anonymous self-assessment surveys and obtained 144 responses. Results suggest that students who do a reproduction project place more value on scientific reproductions and become more critical thinkers. Students and AI researchers agree that our online reproduction repository is valuable.

cs.CY↗