arXiv subjects

Giorgos Flouris

Publications and source records attributed to Giorgos Flouris.

10 recordsLinked to original sources

arXiv2026

Are Safety Guarantees in Neural Networks Safe? How to Compute Trustworthy Robustness Certifications

A primary challenge in AI safety is the existence of adversarial examples -- slightly distorted inputs that cause a neural network (NN) to misclassify. To mitigate this problem, recent research focuses on the computation of robustness certifications, which, for a given input, determine the largest distortion the input may receive without breaking the network's prediction. Robustness certifications can be interpreted as an axis-aligned hyper-rectangle (multi-dimensional intervals). Most existing approaches focus on maximizing the certification's volume, but recent intractability results prohibit the computation of volume-optimal certifications in reasonable time. We introduce the apothem measure and show how to compute apothem-optimal certifications in a linear number of calls to a NN verifier (oracle) w.r.t. the input domain's diameter. Moreover, we prove that we cannot have a volume-optimal, oracle-based algorithm, even if we discard the oracle costs. Also, we introduce dual certifications -- an interval including all instances of a class -- thus providing apothem-minimum upper bounds to a robustness certification. Further, we present the ParallelepipedoNN system, which we evaluate on the standard MNIST and Fashion MNIST benchmarks. A preliminary comparison with existing work on the same datasets reveals at least two-fold improvement w.r.t. the minimum edge length.

cs.LG↗

arXiv2026

The Cost of Relaxation: Evaluating the Error in Convex Neural Network Verification

Many neural network (NN) verification systems represent the network's input-output relation as a constraint program. Sound and complete, representations involve integer constraints, for simulating the activations. Recent works convexly relax the integer constraints, improving performance, at the cost of soundness. Convex relaxations consider outputs that are unreachable by the original network. We study the worst case divergence between the original network and its convex relaxations; both qualitatively and quantitatively. The relaxations' space forms a lattice, where the top element corresponds to a full relaxation, with every neuron linearized. The bottom element corresponds to the original network. We provide analytical upper and lower bounds for the $\ell_\infty$-distance between the fully relaxed and original outputs. This distance grows exponentially, w.r.t. the network's depth, and linearly w.r.t. the input's radius. The misclassification probability exhibits a step-like behavior, w.r.t. input radius. Our results are supported by experiments on MNIST, Fashion MNIST and random networks.

cs.LG↗

arXiv2026

Interval Certifications for Multilayered Perceptrons via Lattice Traversal

In this work we present a rigorous theoretical framework to a foundational problem of AI safety, namely adversarial robustness. In particular, we show that the adversarial robustness problem can be reduced to a lattice traversal problem. Each element of this lattice corresponds to an interval, i.e., an axis-aligned hyper-rectangle, containing an input point $\mathbf{x}$. Consider a multilayered perceptron classifier (MLP). An interval $I$ constitutes a sound certification if $\mathbf{x} \in I$ and $\mathbf{x}$ can be freely perturbed in $I$ without changing the MLP's prediction. Complementarily, an interval $I$ constitutes a complete certification if $\mathbf{x} \in I$ and when $\mathbf{x}$ moves outside of $I$ the MLP's prediction is guaranteed to change. While the sound certification problem corresponds to the well-studied adversarial robustness, complete certifications have not been examined in the literature. We develop lattice traversal operators, which we apply in a refine & verify iterative scheme. Using formal MLP verifiers, sound maximality and complete minimality are guaranteed. Moreover, we examine objective optimization problems. There we discover some interesting asymmetries. For complete certifications, the minimum solution is obtained in polynomial oracle calls. This does not hold for sound certifications, where we prove strong intractability results. Additionally, we examine optimization problems in symmetric intervals (i.e., $\ell_\infty$-spheres), where we provide logarithmic algorithms. Finally, we present an empirical evaluation, using the novel ParallelepipedoNN system.

cs.AI↗

arXiv2025

Modeling and Managing Temporal Obligations in GUCON Using SPARQL-star and RDF-star

In the digital age, data frequently crosses organizational and jurisdictional boundaries, making effective governance essential. Usage control policies have emerged as a key paradigm for regulating data usage, safeguarding privacy, protecting intellectual property, and ensuring compliance with regulations. A central mechanism for usage control is the handling of obligations, which arise as a side effect of using and sharing data. Effective monitoring of obligations requires capturing usage traces and accounting for temporal aspects such as start times and deadlines, as obligations may evolve over times into different states, such as fulfilled, violated, or expired. While several solutions have been proposed for obligation monitoring, they often lack formal semantics or provide limited support for reasoning over obligation states. To address these limitations, we extend GUCON, a policy framework grounded in the formal semantics of SPAQRL graph patterns, to explicitly model the temporal aspects of an obligation. This extension enables the expressing of temporal obligations and supports continuous monitoring of their evolving states based on usage traces stored in temporal knowledge graphs. We demonstrate how this extended model can be represented using RDF-star and SPARQL-star and propose an Obligation State Manager that monitors obligation states and assess their compliance with respect to usage traces. Finally, we evaluate both the extended model and its prototype implementation.

cs.CR↗

arXiv2024

Adaptation Procedure in Misinformation Games

We study interactions between agents in multi-agent systems, in which the agents are misinformed with regards to the game that they play, essentially having a subjective and incorrect understanding of the setting, without being aware of it. For that, we introduce a new game-theoretic concept, called misinformation games, that provides the necessary toolkit to study this situation. Subsequently, we enhance this framework by developing a time-discrete procedure (called the Adaptation Procedure) that captures iterative interactions in the above context. During the Adaptation Procedure, the agents update their information and reassess their behaviour in each step. We demonstrate our ideas through an implementation, which is used to study the efficiency and characteristics of the Adaptation Procedure.

cs.GT↗

arXiv2023

Noisy Games: A Study on the Effect of Noise on Game Specifications

We consider misinformation games, i.e., multi-agent interactions where the players are misinformed with regards to the game that they play, essentially having an \emph{incorrect} understanding of the game setting, without being aware of their misinformation. In this paper, we introduce and study a new family of misinformation games, called Noisy games, where misinformation is due to structured (white) noise that affects additively the payoff values of players. We analyse the general properties of Noisy games and derive theoretical formulas related to ``behavioural consistency'', i.e., the probability that the players behaviour will not be significantly affected by the noise. We show several properties of these formulas, and present an experimental evaluation that validates and visualises these results.

cs.GT↗

arXiv2022

Linguistic Cues of Deception in a Multilingual April Fools' Day Context

In this work we consider the collection of deceptive April Fools' Day(AFD) news articles as a useful addition in existing datasets for deception detection tasks. Such collections have an established ground truth and are relatively easy to construct across languages. As a result, we introduce a corpus that includes diachronic AFD and normal articles from Greek newspapers and news websites. On top of that, we build a rich linguistic feature set, and analyze and compare its deception cues with the only AFD collection currently available, which is in English. Following a current research thread, we also discuss the individualism/collectivism dimension in deception with respect to these two datasets. Lastly, we build classifiers by testing various monolingual and crosslingual settings. The results showcase that AFD datasets can be helpful in deception detection studies, and are in alignment with the observations of other deception detection works.

cs.CL↗

arXiv2021

Deception detection in text and its relation to the cultural dimension of individualism/collectivism

Deception detection is a task with many applications both in direct physical and in computer-mediated communication. Our focus is on automatic deception detection in text across cultures. We view culture through the prism of the individualism/collectivism dimension and we approximate culture by using country as a proxy. Having as a starting point recent conclusions drawn from the social psychology discipline, we explore if differences in the usage of specific linguistic features of deception across cultures can be confirmed and attributed to norms in respect to the individualism/collectivism divide. We also investigate if a universal feature set for cross-cultural text deception detection tasks exists. We evaluate the predictive power of different feature sets and approaches. We create culture/language-aware classifiers by experimenting with a wide range of n-gram features based on phonology, morphology and syntax, other linguistic cues like word and phoneme counts, pronouns use, etc., and token embeddings. We conducted our experiments over 11 datasets from 5 languages i.e., English, Dutch, Russian, Spanish and Romanian, from six countries (US, Belgium, India, Russia, Mexico and Romania), and we applied two classification methods i.e, logistic regression and fine-tuned BERT models. The results showed that our task is fairly complex and demanding. There are indications that some linguistic cues of deception have cultural origins, and are consistent in the context of diverse domains and dataset settings for the same language. This is more evident for the usage of pronouns and the expression of sentiment in deceptive language. The results of this work show that the automatic deception detection across cultures and languages cannot be handled in a unified manner, and that such approaches should be augmented with knowledge about cultural differences and the domains of interest.

cs.CL↗

arXiv2017

On Measuring Bias in Online Information

Bias in online information has recently become a pressing issue, with search engines, social networks and recommendation services being accused of exhibiting some form of bias. In this vision paper, we make the case for a systematic approach towards measuring bias. To this end, we discuss formal measures for quantifying the various types of bias, we outline the system components necessary for realizing them, and we highlight the related research challenges and open problems.

cs.DB↗

arXiv2015

A Flexible Framework for Defining, Representing and Detecting Changes on the Data Web

The dynamic nature of Web data gives rise to a multitude of problems related to the identification, computation and management of the evolving versions and the related changes. In this paper, we consider the problem of change recognition in RDF datasets, i.e., the problem of identifying, and when possible give semantics to, the changes that led from one version of an RDF dataset to another. Despite our RDF focus, our approach is sufficiently general to engulf different data models that can be encoded in RDF, such as relational or multi-dimensional. In fact, we propose a flexible, extendible and data-model-independent methodology of defining changes that can capture the peculiarities and needs of different data models and applications, while being formally robust due to the satisfaction of the properties of completeness and unambiguity. Further, we propose an ontology of changes for storing the detected changes that allows automated processing and analysis of changes, cross-snapshot queries (spanning across different versions), as well as queries involving both changes and data. To detect changes and populate said ontology, we propose a customizable detection algorithm, which is applicable to different data models and applications requiring the detection of custom, user-defined changes. Finally, we provide a proof-of-concept application and evaluation of our framework for different data models.

cs.DB↗