Papers by Petros Stefaneas
This paper focuses on the security of electronic mail, using machine learning algorithms. Spam em... more This paper focuses on the security of electronic mail, using machine learning algorithms. Spam email is unwanted messages, usually commercial, sent to a large number of recipients. In this work, an algorithm for the detection of spam messages with the aid of machine learning methods is proposed. The algorithm accepts as input text email messages grouped as benevolent ("ham") and malevolent (spam) and produces a text file in csv format. This file then is used to train a bunch of ten Machine Learning techniques to classify incoming emails into ham or spam. The following Machine Learning techniques have been tested: Support Vector Machines, k-Nearest Neighbour, Naïve Bayes, Neural Networks, Recurrent Neural Networks, Ada Boost, Random Forest, Gradient Boosting, Logistic Regression and Decision Trees. Testing was performed using two popular datasets, as well as a publicly available csv file. Our algorithm is written in Python and produces satisfactory results in terms of accuracy, compared to state-of-the-art implementations. In addition, the proposed system generates three output files: a csv file with the spam email IP addresses (of originating email servers), a map with their geolocation, as well as a csv file with statistics about the countries of origin. These files can be used to update existing organisational filters and blacklists used in other spam filters.
A cohesive crack model is proposed to describe strain localization for the materials where strain... more A cohesive crack model is proposed to describe strain localization for the materials where strain-hardening is not prevailing over strain-softening (geomaterials, concrete-like materials, ceramics, etc.). Such a model is able to predict the size effects of fracture mechanics, i.e., the transition from ductile to brittle structure behaviour by increasing the size scale and keeping the geometrical shape unchanged. Whereas for Mode I, only untieing of the finite element nodes is applied to simulate crack growth, for Mixed Mode interelement crack propagation a topological variation is required at each step. In the case of four point shear testing, the load vs. deflection diagrams reveal snap-back instability for large sizes. By increasing the specimen sizes, such instability tends to reproduce the classical LEFM instability, predicted by the Maximum Circumferential Stress Criterion . Experimentally, the fracture toughness parameter of concrete appears to be unique and represented by the Mode I fracture energy Gp or the stress-intensity factor Kic even for Mixed Mode problems.
EΦΑΡΜΟΓEΣ ΤΗΣ ΛΟΓΙΚΗΣ ΣΤΗΝ ΠΛΗΡΟΦΟΡΙΚΗ [EΘΝΙΚΟ ΜEΤΣΟΒΙΟ ΠΟΛΥΤEΧΝEΙΟ, ΣΧΟΛΗ EΦΑΡΜΟΣΜEΝΩΝ ΜΑΘΗΜΑΤΙΚ... more EΦΑΡΜΟΓEΣ ΤΗΣ ΛΟΓΙΚΗΣ ΣΤΗΝ ΠΛΗΡΟΦΟΡΙΚΗ [EΘΝΙΚΟ ΜEΤΣΟΒΙΟ ΠΟΛΥΤEΧΝEΙΟ, ΣΧΟΛΗ EΦΑΡΜΟΣΜEΝΩΝ ΜΑΘΗΜΑΤΙΚΩΝ ΚΑΙ ΦΥΣΙΚΩΝ EΠΙΣΤΗΜΩΝ, ΤΜΗΜΑ EΦΑΡΜΟΣΜEΝΩΝ ΜΑΘΗΜΑΤΙΚΩΝ ΚΑΙ ΦΥΣΙΚΩΝ EΠΙΣΤΗΜΩΝ]
Journal of Logic and Computation, Jun 6, 2023
In this paper, we address modelling program development by multiple different programmers (or pro... more In this paper, we address modelling program development by multiple different programmers (or programming teams) who work in different settings (programming languages or reasoning frameworks) but follow a common specification. After consideration of some philosophical issues regarding program development, including its similarities and dissimilarities with scientific theorizing, we extend a logical framework built to describe scientific theorizing in relativist settings: our extensions add the ability of reasoning about programs, the iterative process of their generation and their specifications. We are thus able to use the framework in order to make a first step towards examining at an abstract level the problem of translatability between the programs produced by program generators, by defining a notion of translation between the outputs of program generators and proving that there is such a (trivial) translation when two program generators follow the same specification reliably (in a specific sense of reliability).
arXiv (Cornell University), Nov 2, 2022
In this paper, we address program development by multiple different programmers (or programming t... more In this paper, we address program development by multiple different programmers (or programming teams), each working in different settings (programming languages or reasoning frameworks), but following a common specification; in particular, we examine at an abstract level the problem of translatability between their produced programs. To this end, after consideration of some philosophical issues regarding program development, including its similarities and dissimilarities with scientific theorising, we extend a logical framework built to describe scientific theorising in relativist settings: our extensions add the ability of reasoning about programs, the iterative process of their generation, and their specifications. We are thus able to define a notion of translation between the outputs of program generators and prove that there is a (trivial) such translation when two program generators follow the same specification reliably (in a specific sense of reliability).
Lecture Notes in Computer Science, 2018
Philippou and Kouzapas have proposed a privacy-related framework, consisting of (i) a variant of ... more Philippou and Kouzapas have proposed a privacy-related framework, consisting of (i) a variant of the \(\pi \)-calculus, called Privacy Calculus, that describes the interactions of processes, (ii) a privacy policy language, (iii) a type system that serves to check whether Privacy Calculus processes respect privacy policies. We present an executable implementation of (a version of) it in the programming/specification language Maude: we give an overview of the framework, outline the key aspects of its implementation, and offer a simple example of how the implementation can be used.
arXiv (Cornell University), May 3, 2023
This whitepaper has been produced by members of the COST Action 18230-Interactive Narrative Desig... more This whitepaper has been produced by members of the COST Action 18230-Interactive Narrative Design for Complexity Representation (INDCOR-https://indcor.eu) INDCOR White Paper 2: Interactive Narrative Design for Representing Complexity Executive Summary This white paper was written by the members of the Work Group focusing on design practices of the COST Action 18230-Interactive Narrative Design for Complexity Representation (INDCOR, WG1). It presents an overview of Interactive Digital Narratives (IDNs) design for complexity representations through IDN workflows and methodologies, IDN authoring tools and applications. It provides definitions of the central elements of the IDN alongside its best practices, designs and methods. Finally, it describes complexity as a feature of IDN, with related examples. In summary, this white paper serves as an orienting map for the field of IDN design, understanding where we are in the contemporary panorama while charting the grounds of their promising futures.
We propose the extension of previous work done on “the Web as a tool for proving” [2] to suggest ... more We propose the extension of previous work done on “the Web as a tool for proving” [2] to suggest a basis for a meta-theory for post-truth. We believe that such a theory cannot be based on the current philosophical theories about truth such as the correspondence theory, the coherence theory, the pragmatic theory or the consensus theory [8]. Neither can be based on the constructivist approach that truths in general are social constructions. “Post-truths” are often considered in the literature as real existing entities [7]; the only aspect that takes place in the real world is the processes of assigning truth-values to assertions. Our view is that the truth-assignment processes involve at least two agents – the claim initiator and the truth interpreter - which can be persons, groups of persons or even machines. The same kind of approach applies for Web-based proving which occurs at a particular time and place and involves particular people, some of them can act as administrators. Underlying the truth-assignment processes and the proving processes over the Web, we have some kind of social event. The nature of the Web allows the participation of people who have not necessarily particular skills as members of a certain group. Web proving has been studied [2] as a particular type of Goguen's proof events [3][4]. Proof events, are social events that involve particular persons that form social groups of experts with particular knowledge and skills. These groups are open and have no internal hierarchical structure but they usually have at least one administrator who acts as an overseer of the correctness of the proving processes. In any case, a proof event apart from involving specialists, involves mediating objects such as spoken words, data, videos, scientific papers, etc. Web-based proof-events have a social component, communication medium, prover-interpreter interaction, interpretation process, understanding and validation, historical component, and styles. By truth-interpretation we understand the determination of the definition or meaning of the signs that are fixed by the language or semiotic code of communication used for the claim initiation or for what is thought as truth- assignment. Interpretation is an active process of interactive nature, as allowed by the open Web architecture [1][2]. The concept of post-truth has received a lot of publicity the last few years [7]. It refers to situations that claims can be accepted on the basis of beliefs or emotions and not of real facts [7] [8] [9]. As a concept, it originates from the study of misinformation and in particular fake news [6] [10]. However, its formal epistemological definition is an open question. In our view we need to re-approach the process of truth-assignment within the Web technology, in order to develop a proper theory on post truth based on agents within a social environment. Many social agents that can contribute towards this goal. Still, two of them play the most crucial part in this direction: the claim initiator and the truth-interpreter, both seen as integral parts in the social approach to truth-assignment processes. This theory needs to be accompanied by the necessary socio-theoretical approach regarding the concepts of correctness and proving. Mathematical logic uses mainly mechanical methods, such as rules of inference and validation, and in a sense considers that most of the truth-related components of a logical system can be constructed as finite sequences of such rules. This approach leaves out the social dimension of the truth-assignment process as well as the social dimension of proofs, including their histories, attempts to arrive at a true conclusion, motivations, misleading interpretations, etc. Truth-assignments have much greater diversity than what most of logicians would easily accept. Any post-truth statement depends much on the processes used for the truth assignment. Depending on which processes are used, the output differs. To conclude that “an assertion is true”, a broader notion of correctness is desirable. In the literature, correctness refers to whether or not an argument or proof follows a logical path from premises to conclusions [12]. We suggest that, apart from this logical correctness, we need a rule-based and a morally-based conception of correctness to encapsulate accurately the social aspects. Correctness of an action over the Web needs to re-defined as an action that is not only logically acceptable but complies with these kinds of social norm [11]. The open architecture of the Web, facilitates certain social behaviors and prevents others, thus it provides novel features of the truth assignment process far beyond the traditional approach to argumentation within a natural language.
In recent years, deep neural networks, especially those exhibiting synergistic properties, have b... more In recent years, deep neural networks, especially those exhibiting synergistic properties, have been at the cutting edge of image processing, producing very good results. So far, they have been able to successfully address issues of classification and recognition of objects depicted on images. In this paper, a novel idea is presented, where images of chemical structures are used as input information in deep learning neural network architectures aiming at the generation of Quantitative Structure Activity Relationship (QSAR) models, i.e. models that predict properties, activities or adverse effects of chemicals. The proposed method was applied to a case study of particular interest, which is the predic
Journal of Engineering Research and Sciences, May 1, 2022
JavaScript Object Notation (JSON) is an open standard data format that is used widely across the ... more JavaScript Object Notation (JSON) is an open standard data format that is used widely across the internet as means of exchanging structured data due to its low overhead. While originally created in the early 2000s, it has only gained standard status in 2013 and then again in 2017 with a new version that focused more on security and interoperability. In this paper the authors present a different specification of the JSON standard that relies on algebraic formal methods and provides certain benefits over a regular natural language specification. This specification can also function as a schema that can attest a JSON data document's compliance to its blueprint. The absorption of Formal Specification methods by the industry happens at a very slow pace, mostly because there is little incentive to tread into a fairly unknown territory. Notwithstanding this reluctance, the authors encourage the usage of Formal Specification techniques to the specifications of open standards; Formal specifications are more succinct, less ambivalent, consistent to the standard, reusable as they support module inheritance and can be executable. The process of designing new Standards can benefit from Formal Specifications as the resulting specification i) is more tangible; ii) allows a thorough and clear understanding of the standard and also iii) allows property checking and property verification.
Springer proceedings in mathematics & statistics, 2017
We collect results related to generic constructions and generic limits for semantic and syntactic... more We collect results related to generic constructions and generic limits for semantic and syntactic cases. It is considered both by pure model theory approach and by the institutional approach.
International Journal of Web Information Systems, Nov 7, 2016
Purpose Reactive rules are used for programming rule-based Web agents, which have the ability to ... more Purpose Reactive rules are used for programming rule-based Web agents, which have the ability to detect events and respond to them automatically and can have complex structure and unpredictable behavior. The aim of this paper is to provide an appropriate formal framework for analyzing such rules. Design/methodology/approach To achieve this goal, the authors give two alternative semantics for the basic reactive rules’ families which allow us to specify reactive rule-based agents and verify their intended behavior. The first approach expresses the functionality of production and event condition action rules in terms of equations, whereas the second methodology is based in the formalism of rewriting logic. Both semantics can be expressed within the framework of CafeOBJ algebraic specification language, which then offers the verification support and have their advantages and downsides. Findings The authors report on experiences gained by applying those methodologies in a reactive rule-based system and compare the two methodologies. Originality/value Finally, the authors demonstrate a tool that translates a set of reactive rules into CafeOBJ rewrite rules, thus making the verification of reactive rules possible for inexperienced users.
In this paper we advocate for applying formal specification techniques to open standards’ specifi... more In this paper we advocate for applying formal specification techniques to open standards’ specifications by creating an algebraic schema for JSON data file format, an open standard file format that has been very popular when it comes to the exchange of structured data. Integrating formal specification methods and standards allows for better standard design as i) the results are more tanbigle; ii) it enables a deeper understanding of the standard; iii) enforces clarity into the specification, and also iv) provides extended property checking/verification capabilities, particularly if they use any algebraic specification language. The resulting specification can be used as a JSON schema that can validate JSON documents.
Сибирские электронные математические известия, Nov 29, 2019
We define and study lattices in generative classes associated with generic structures. It is show... more We define and study lattices in generative classes associated with generic structures. It is shown that these lattices can be non-distributive and, moreover, arbitrary enough. Heights and wights of the lattices are described. A model-theoretic criterion for the linear ordering is proved and these linear orders are described. Boolean algebras generated by the considered lattices are also described.
Lecture notes in business information processing, 2019
This paper presents an early version of a decision-making “eco” system. We refer to it as an “eco... more This paper presents an early version of a decision-making “eco” system. We refer to it as an “eco” system because it is primarily based on mathematical logic and combines concepts and principles from the fields of statistics, decision theory, artificial intelligence and modeling of human behavior. The primary goal of the proposed approach is to address errors that occur resulting from the misuse of statistical methods. In practice, such errors often occur either owning to the use of inappropriate statistical methods or wrong interpretations of results. The proposed approach relies on the LPwNF (Logic Programming without Negation as Failure) framework of non-monotonic reasoning as provided by Gorgias. The proposed system enables automatic selection of the appropriate statistical method, based on the characteristics of the problem and the sample. The expected impact could be twofold: it can enhance the use of statistical systems like R and, combined with a Java-based interface to Gorgias, make non-monotonic reasoning easy to use in the proposed context.
International Journal of Software Engineering and Knowledge Engineering, Sep 1, 2015
Context-aware and self adaptive systems have become very popular during the last decade. As these... more Context-aware and self adaptive systems have become very popular during the last decade. As these technologies are increasingly used in the development of critical applications, their behavior should be extensively analyzed. While formal methods provide a wide range of techniques for reasoning about software systems, addressing formally the requirements of context-aware adaptive systems in a consistent way remains a challenge. To this end, in this paper we present an algebraic framework for their formal specification using Observational Transition Systems (OTSs) specified in the CafeOBJ algebraic specification language. This approach permits the verification of the design of such systems, and can be an effective approach to obtaining verified context-aware software. We apply the proposed framework to the modeling of a context-aware adaptive traffic monitoring system and use theorem proving techniques to prove safety properties for that system.
Synthese Library, 2019
Selected historical and contemporary aspects of the debate between the syntactic and semantic vie... more Selected historical and contemporary aspects of the debate between the syntactic and semantic views of scientific theories are presented. The theory of institutions is an abstract model theory approach which allows representation of both the syntactic and semantic views of scientific theories in a way which is free from dependence upon particular formal languages and logics. The standard definition of an institution, and the logical equivalence of the syntactic and the semantic definitions of a theory over an institution are presented. The value of the theory of institutions for formalising computer science theories is reviewed and the wider methodological implications of this approach are considered. Formalising social science theories in the theory of institutions has a number of benefits including enabling approaches which simultaneously require the use of more than one theoretical vocabulary and permitting transitions between different levels of abstraction.
Uploads
Papers by Petros Stefaneas