Skip to main content

Chairi Kiourt

Research and Innovation Centre “Athena”, Institute for Language and Speech Processing, Faculty Member

Followers

46

Following

35

Co-authors

10

Public Views

Dr. Chairi Kiourt is assistant researcher in ATHENA - Research and Innovation Centre in Information, Communication and Knowledge Technologies. Also, since 2018, is Posdoctoral researcher with the Hellenic Open University, School of Science and TechnologyHe has participated in several research programs and has contributed to several scientific publications in international journals and conferences with judges.

less

Mediterranean Archaeology and Archaeometry (MAA)

University of the Aegean

Anestis Koutsoudis

Ioannis Paliokas

Margarita Skamantzari

Georgia Kontogianni

National Technical University of Athens

Andreas Georgopoulos

National Technical University of Athens

Konstantina Siountri

Shenkar College of Engineering and Design

InterestsView All (16)

Uploads

Papers by Chairi Kiourt

Realistic Simulation of Cultural Heritage

IGI Global eBooks, 2020

One of the most challenging problems in the simulation of real environments is to generate worlds... more One of the most challenging problems in the simulation of real environments is to generate worlds that appear realistic and more attractive. It becomes increasingly challenging when the simulated environment focuses on minors (students), because the young generation has high demands on simulation systems due to their experience in computer gaming. Virtual museums are among the most important simulation environments, which present cultural and educational content for everyone. Their purpose is to enrich the users experience by allowing an intuitive interaction with the museum artifacts and to offer knowledge with the most pleasant ways. This paper focuses on the aspects of realistic simulations in the development of virtual 3D environments for Cultural Heritage applications. This study includes aspects regarding some of the most high-tech image effects, applicable artificial intelligence methods, powerful game engines, how real object can be reconstructed realistically and how all those features may be combined to produce realistic, pleasant, productive and educative environments.

Enhanced Virtual Reality Experience in Personalised Virtual Museums

IGI Global eBooks, 2020

Thisarticlefocusesonimportantfactorsinthecreationofenhancedpersonalisedexperiencesin virtualenvir... more Thisarticlefocusesonimportantfactorsinthecreationofenhancedpersonalisedexperiencesin virtualenvironmentsforculturalheritageapplications,especiallythosetargetingvirtualmuseumsand exhibitions.Someofthemostimportantfactorsrelatingtopersonalisedvirtualmuseumsthatrelate tointelligentcontentandusermodellinginvirtualenvironmentsarebeinghighlightedanddiscussed. After an extensive review of the current trends in the domain, the article presents a generalised frameworkforthedevelopmentofthenextgenerationenhancedVRexperiencesinpersonalised virtualmuseums.Thisframework,whichnaturallysurfacesfromthedomain,wasputtothetestin thedevelopmentoftheDynaMusplatformandtwocasestudiesbasedonthisplatformarereferenced andcommentedtosupportsuchanapproach.Thisconceptcanserveasthegeneralframeworkfor developingenhancedpersonalisedvirtualenvironmentsforculturalheritageapplications.

Augmenting existing food image datasets with Greek dishes

Zenodo (CERN European Organization for Nuclear Research), Oct 21, 2021

1 Extensive study has been done in the framework of the relevant Greek national project GRE-Taste... more

Collective artificial intelligence and personalized educational interactions in learning new skills

Η Κοινωνική Μάθηση (Social Learning) αποτελεί μια νέα, σημαντική περιοχή έρευνας στον τομέα της Τ... more Η Κοινωνική Μάθηση (Social Learning) αποτελεί μια νέα, σημαντική περιοχή έρευνας στον τομέα της Τεχνητής Νοημοσύνης (Artificial Intelligence) και συγκεκριμένα στην μελέτη των μηχανισμών μάθησης των πολυπρακτορικών συστημάτων (Multi-Agent Systems) παιχνιδιού. Επίσης, θεωρείτε ένας σημαντικός άξονας μελέτης και εφαρμογών για διάφορους επιστημονικούς τομείς, όπως η κοινωνιολογία, η οικονομία και το μάρκετινγκ. Οι κοινωνικοί οργανισμοί (Social Organizations), βασιζόμενοι στα συστήματα προσομοίωσης με πράκτορες, καθώς και η μελέτη/διερεύνηση των στρατηγικών/τεχνικών εκπαίδευσης και μάθησης των πρακτόρων, εμπνεύστηκαν από την ικανότητα των ανθρώπων να μαθαίνουν από κοινωνικά περιβάλλοντα που είναι πλούσια σε κρυφές πληροφορίες και σε αλληλεπιδράσεις μεταξύ οντοτήτων (π.χ. ανθρώπων). Αυτός ο πλούτος του περιβάλλοντος, είναι μια πηγή πολυπλοκότητας όπου ένας μαθητευόμενος πρέπει να μπορεί να πλοηγηθεί ορθά και αποτελεσματικά σύμφωνα με τον στόχο του. Η εισαγωγή της κοινωνικής δυναμικής στα πολυπρακτορικά περιβάλλοντα, θεωρείται ως μια πολλά υποσχόμενη προσέγγιση για την προσομοίωση ρεαλιστικών συμπεριφορών παιχνιδιού, συμπεριφορές που μοιάζουν με ανθρώπινες, καθώς και για την ανάδειξη της σημαντικότητας της Συλλογικής Γνώσης (Collective Knowledge).Η διατριβή αυτή, στοχεύει στην έρευνα της απόδοσης των συνθετικών πρακτόρων (Synthetic Agents) σε σενάρια μάθησης και παιξίματος σε παιχνίδια στρατηγικής με μηδενικό άθροισμα (Zero-Sum). Επίσης, αναδεικνύουμε την σημαντικότητα και την ικανότητα των μοντέλων μάθησης με βάση τον αντίπαλο (Opponent Based Learning), που παρουσιάζουν τα ανταγωνιστικά κοινωνικά περιβάλλοντα με στόχο την αποδοτικότερη μάθηση και παίξιμο ενός πράκτορα.Οι συνθετικοί πράκτορες σχεδιάστηκαν και αναπτύχθηκαν με βάση διάφορους συνδυασμούς κάποιων βασικών παραμέτρων μάθησης, όπως η επιλογή μεταξύ εξερεύνησης και εκμετάλλευσης γνώσης, ταχύτητα μάθησης κ.α.. Στη συνέχεια, οι συνθετικοί πράκτορες αλληλεπιδρούν σε μεγάλους αριθμούς παιχνιδιών και τα πειραματικά δεδομένα αναλύονται για τον προσδιορισμό των ομαδοποιήσεων τους, που παρουσιάζουν σημαντικές συσχετίσεις μεταξύ των παραμέτρων μάθησης και την τελική κατάταξη των επιδόσεων τους.Επιπλέον, εξετάζουμε πώς οι συνθετικοί πράκτορες αλληλεπιδρούν σε κοινωνικά περιβάλλοντα, χρησιμοποιώντας διάφορες στρατηγικές εκπαίδευσης εναντίον διαφόρων αντιπάλων (πράκτορες με διαφορετικά χαρακτηριστικά μάθησης και παιξίματος). Αυτός ο πειραματισμός εκπαίδευσης και παιξίματος, αναδεικνύει πως η ποιότητα του παιξίματος εξαρτάται περισσότερο από τη ορθή παραμετροποίηση του μηχανισμού μάθησης παρά από την εμπειρία.Εξετάζουμε επίσης, την πρόοδο/εξέλιξη της μάθησης των άπειρων πρακτόρων σε κοινωνικά περιβάλλοντα ανταγωνιστικών παιχνιδιών, στοχεύοντας στον προσδιορισμό της επίδρασης ενός έμπειρου αντιπάλου σε έναν αρχάριο πράκτορα. Η διερεύνηση των επιπτώσεων της πολυπλοκότητας του περιβάλλοντος στη συμπεριφορά παιξίματος και μάθησης των συνθετικών πρακτόρων, αναδεικνύει τον τρόπο με τον οποίο ένας αποτελεσματικός παίκτης πρέπει να προσαρμόσει το προφίλ μάθησης και παιξίματος του, για να διατηρήσει ένα συγκεκριμένο προφίλ απόδοσης, όταν η πολυπλοκότητα τους κοινωνικού περιβαλλοντικές μεταβάλλεται.Για την επιβεβαίωση των πειραμάτων αποτελεσμάτων, διεξάγουμε μια αξιολόγηση, μεγάλης κλίμακας, των υπαρχουσών μεθόδων αξιολόγησης της επίδοσης των παικτών (Elo και Glicko), εφαρμοσμένες σε πολυπρακτορικά συστήματα, όπου τονίζεται μια ασυνέπεια (ασυμφωνία) ως προς τον τρόπο με τον οποίο οι προαναφερθέν μέθοδοι αξιολογούν της επιδόσεις των συνθετικών πρακτόρων. Στην συνέχεια, προτείνουμε μία νέα προσέγγιση αξιολόγησης της επίδοσης των πρακτόρων. Η προσέγγιση αυτή, μπορεί να θεωρηθεί ως βάση για την ανάπτυξη μεθόδων αξιολόγησης της επίδοσης παικτών για πολυπρακτορικά συστήματα. Με ένα μεγάλο αριθμό πειραμάτων αποδεικνύεται η αποτελεσματικότητά της προσέγγισης που προτάχθηκε.Επίσης, στα πλαίσια της διατριβής αυτής, παρουσιάζεται μία μέθοδος επιλογής αντιπάλου από μία ομάδα διαφορετικών πρακτόρων, με σκοπό την αποτελεσματικότερη μάθηση και παίξιμο. Μελετήθηκε ένας συνθετικός πράκτορας (με προφίλ καλού παίκτη) όπου πειραματίζεται σε παιχνίδια στρατηγικής με διάφορες αλληλουχίες διαφορετικών αντιπάλων (πράκτορες με διαφορετικά χαρακτηριστικά παιξίματος και επιδόσεις) για τον προσδιορισμό της αλληλουχίας που θα του προσφέρει την &quot;αποτελεσματικότερη μάθησης&quot;. Τα αποτελέσματα των πειραμάτων, δείχνουν ότι η αποτελεσματικότερη πρόοδος και η σταθερότερη εξέλιξη ενός πράκτορα, προκύπτει όταν αυτός αντιμετωπίζει αρχικά αντιπάλους με χειρότερο προφίλ παιξίματος από το δικό του και σταδιακά επιλέγει αντιπάλους με καλύτερο προφίλ παιξίματος. Επίσης, οι μελέτες δείχνουν πως ένας συνθετικός πράκτορας με σωστά διαμορφωμένο/παραμετροποιημένο μηχανισμό μάθησης αποδίδει καλύτερα όταν αντιμετωπίζει λιγότερο ευνοϊκά διαμορφωμένους πράκτορες.Η προσομοίωση των κοινωνικών περιβαλλόντων, όπως στα προαναφερθέντα πειράματα, απαιτεί τεράστιους υπολογιστικούς πόρους. Για τον λόγο αυτό τα πειράματα της…

Deep Learning-Based Black Spot Identification on Greek Road Networks

Data, Jun 16, 2023

This article is an open access article distributed under the terms and conditions of the Creative... more

Fat calculation from raw-beef-steak images through machine learning approaches: an end-to-end pipeline

Proceedings of the 26th Pan-Hellenic Conference on Informatics, Nov 25, 2022

eTHOS: a web-based tax morale system

International Journal of Electronic Governance

In this paper, an innovative web-based cross-platform information system, named eTHOS, is introdu... more In this paper, an innovative web-based cross-platform information system, named eTHOS, is introduced allowing citizens to either report various complaints, in real-time, about tax evasion incidents, or to highlight good tax behavior practices. This is achieved through the creation of interactive maps that highlight geospatial financial factors for the general public. The main purpose of the system is to capture specific information which will be used in order to understand the various tax morale approaches that lead citizens to file a report (complaint or suggestions for reward) based on their interaction with businesses. Furthermore, the new proposed system aims to gain an operational role as the first citizen oriented digital tool against tax evasion. The contribution of eTHOS to the economy could be important, since it could lead to the containment of tax evasion and the decrease of the VAT Gaps.

Semantic Image Segmentation with Deep Learning for Vine Leaf Phenotyping

IFAC-PapersOnLine

Plant phenotyping refers to a quantitative description of the plant's properties, however in imag... more Plant phenotyping refers to a quantitative description of the plant's properties, however in image-based phenotyping analysis, our focus is primarily on the plant's anatomical, ontogenetical and physiological properties. This technique reinforced by the success of Deep Learning in the field of image based analysis is applicable to a wide range of research areas making high-throughput screens of plants possible, reducing the time and effort needed for phenotypic characterization. In this study, we use Deep Learning methods (supervised and unsupervised learning based approaches) to semantically segment grapevine leaves images in order to develop an automated object detection (through segmentation) system for leaf phenotyping which will yield information regarding their structure and function. In these directions we studied several deep learning approaches with promising results as well as we reported some future challenging tasks in the area of precision agriculture. Our work contributes to plant lifecycle monitoring through which dynamic traits such as growth and development can be captured and quantified, targeted intervention and selective application of agrochemicals and grapevine variety identification which are key prerequisites in sustainable agriculture.

Συλλογική τεχνητή νοημοσύνη και προσωποποιημένες εκπαιδευτικές αλληλεπιδράσεις στην εκμάθηση νέων δεξιοτήτων

Social Learning has been an exciting area in Artificial Intelligence for researching learning mec... more Social Learning has been an exciting area in Artificial Intelligence for researching learning mechanisms for in multi-agent game playing systems, besides being a useful concept in diverse science areas, like sociology, economics and marketing. The agent-based simulation of social organizations, based on the investigation of agents’ training and learning tactics and strategies, has been inspired by the ability of humans to learn from social environments which are rich in interactions between peers and partial or hidden information. Such richness is a source of complexity that an effective learner has to be able to navigate. Introducing socially inspired dynamics in multi-agent environments is perceived as a promising approach to simulate realistic human-like playing behaviors and highlight the importance of collectively gained knowledge.In this thesis, we first investigate the performance of synthetic agents in playing and learning scenarios in a turn-based zero-sum game and we highl...

Multi-agents based virtual environments for cultural heritage

2017 XXVI International Conference on Information, Communication and Automation Technologies (ICAT)

This paper focuses on the integration of artificial intelligence methods in multi-agent virtual e... more This paper focuses on the integration of artificial intelligence methods in multi-agent virtual environments for cultural heritage applications and especially virtual museums and exhibitions. Multi-agent systems are considered here as being autonomous social organizations capable of developing dynamic personalized virtual environments. In this respect, some of the most common artificial intelligence methods in intelligent virtual environments are discussed. Three different sub-domains of intelligent virtual environments are identified and presented, aiming at an enhanced virtual reality experience. The paper concludes by highlighting the importance of artificial intelligence applications in cultural heritage, and the dynamics of the virtual museums with personalized content, based on the profile of the user and the need of the case.

Sustainable Ecotourism through Cutting-Edge Technologies

Sustainability, 2022

Tourism is a phenomenon that dates back to ancient times. Ancient Greek philosophers recognised, ... more Tourism is a phenomenon that dates back to ancient times. Ancient Greek philosophers recognised, adopted, and promoted the concept of rest-based tourism. Ecotourism is a particular type of tourism that connects with activities that take place in nature, without harming it, along with the herbal and animal wealth. According to estimates, the global ecotourism industry is currently booming due to various reasons, and it is becoming an important factor of sustainable regional development. This article presents the vision, work, and outcomes of project AdVENt, a project focusing natively in sustainable ecotourism through natural science and technological innovation. AdVENt’s study area includes the National Parks of Oiti (or Oeta) and Parnassus in Central Greece, where there is a remarkable native flora with a high endemism rate integrated with areas of cultural value and national and European hiking routes and paths of varying difficulty.

Deep Learning Platforms in Food Recognition

24th Pan-Hellenic Conference on Informatics, 2020

Image content prediction with novel deep learning approaches is a hot research topic for many sci... more Image content prediction with novel deep learning approaches is a hot research topic for many scientific disciplines. Image-based food types recognition and their ingredients is a particularly challenging task, since food dishes are typically deformable objects, usually including complex semantics, which makes the task of defining their structure very difficult. In this paper we introduce a novel web-based system that exploits commercial deep learning based platforms to predict image content for food recognition. In addition to that, the system combines the individual predictions of the platforms to produce more accurate results. Also, we provide a short assessment of the chosen platforms to determine the most efficient one, in the domain of food recognition. The paper is concluded by highlighting the key features of platforms and the advantages of the new presented system.

Personalized Interactive Edutainment in Extended Reality (XR) Laboratories

ERCIM News, 2020

AnCoins: Image-Based Automated Identification of Ancient Coins Through Transfer Learning Approaches

Pattern Recognition. ICPR International Workshops and Challenges, 2021

Metis: Multi-Agent Based Crisis Simulation System

With the advent of the computational technologies (Graphics Processing Units - GPUs) and Machine ... more With the advent of the computational technologies (Graphics Processing Units - GPUs) and Machine Learning, the research domain of crowd simulation for crisis management has flourished. Along with the new techniques and methodologies that have been proposed all those years, aiming to increase the realism of crowd simulation, several crisis simulation systems/tools have been developed, but most of them focus on special cases without providing users the ability to adapt them based on their needs. Towards these directions, in this paper, we introduce a novel multi-agent-based crisis simulation system for indoor cases. The main advantage of the system is its ease of use feature, focusing on non-expert users (users with little to no programming skills) that can exploit its capabilities a, adapt the entire environment based on their needs (Case studies) and set up building evacuation planning experiments with some of the most popular Reinforcement Learning algorithms. Simply put, the syste...

Crowd simulation for crisis management: The outcomes of the last decade

Machine Learning with Applications, 2020

The last few decades, crowd simulation for crisis management is highlighted as an important topic... more The last few decades, crowd simulation for crisis management is highlighted as an important topic of interest for many scientific fields. As the continues evolution of computational resources increases, along with the capabilities of Artificial Intelligence, the demand for better and more realistic simulation has become more attractive and popular to scientists. Along those years, there have been published hundreds of research articles and have been created numerous different systems that aim to simulate crowd behaviors, crisis cases and emergency evacuation scenarios. For better outcomes, recent research has focused on the separation of the problem of crisis management, to multiple research sub-fields (categories), such as the navigation of the simulated pedestrians, their psychology, the group dynamics etc. There have been extended research works suggesting new methods and techniques for those categories of problems. In this paper, we propose three main research categories, each one consist of several sub-categories, relying on crowd simulation for crisis management aspects and we present the outcomes of the last decade, focusing mostly on works exploiting multi-agent technologies. We analyze a number of technologies, methodologies, techniques, tools and systems introduced throughout the last years. A comparative review and discussion of the proposed categories is presented towards the identification of the most efficient aspects of the proposed categories. A general framework, towards the future crowd simulation for crisis management is presented based on the most efficient to yield the most realistic outcomes of the last decades. The paper is concluded with some highlights and open questions for future directions.

Rating the skill of synthetic agents in competitive multi-agent environments

Knowledge and Information Systems, 2018

A very effective and promising approach to simulate real-life conditions in multi-agent virtual e... more A very effective and promising approach to simulate real-life conditions in multi-agent virtual environments with intelligent agents is to introduce social parameters and dynamics. Introduction of social parameters in such settings reshapes the overall performance of the synthetic agents, so a new challenge of reconsidering the methods to assess agents' evolution emerges. In a number of studies regarding such environments, the rating of the agents is being considered in terms of metrics (or measures or simple grading) designed for humans, such as Elo and Glicko. In this paper, we present a large-scale evaluation of existing rating methods and a proposal for a new rating approach named Relative Skill-Level Estimator (RSLE), which can be regarded as a base for developing rating systems for multi-agent systems. The presented comparative study highlights an inconsistency in rating synthetic agents in the context described by the widely used methods and demonstrates the efficiency of the new RSLE.

On Structure-From-Motion Application Challenges

International Journal of Computational Methods in Heritage Science, 2017

Numerous software solutions that implement the Structure-from-Motion/Multi-View Stereo (SfM/MVS) ... more Numerous software solutions that implement the Structure-from-Motion/Multi-View Stereo (SfM/MVS) 3D reconstruction approach have been made available during the last decade. These allow the production of high quality in terms of geometry and colour information 3D models with the use of unordered image collections that depict a static scene or object from arbitrary viewpoints. Nowadays SfM/MVS-based 3D reconstruction approaches constitute a popular solution in a variety of applications within the cultural heritage domain. As with all 3D reconstruction approaches it has its limitations and application challenges. In this work, the authors attempt to compose a set of guidelines that are based on the important outcomes of important published works that propose solutions that overcome the various challenges introduced by non-friendly surface types.

Using opponent models to train inexperienced synthetic agents in social environments

2016 IEEE Conference on Computational Intelligence and Games (CIG), 2016

This paper investigates the learning progress of inexperienced agents in competitive game playing... more This paper investigates the learning progress of inexperienced agents in competitive game playing social environments. We aim to determine the effect of a knowledgeable opponent on a novice learner. For that purpose, we used synthetic agents whose playing behaviors were developed through diverse reinforcement learning set-ups, such as exploitation-vs-exploration trade-off, learning backup and speed of learning, as opponents, and a self-trained agent. The paper concludes by highlighting the effect of diverse knowledgeable synthetic agents in the learning trajectory of an inexperienced agent in competitive multiagent environments.

Synthetic learning agents in game-playing social environments

Adaptive Behavior, 2016

This paper investigates the performance of synthetic agents in playing and learning scenarios in ... more This paper investigates the performance of synthetic agents in playing and learning scenarios in a turn-based zero-sum game and highlights the ability of opponent-based learning models to demonstrate competitive playing performances in social environments. Synthetic agents are generated based on a variety of combinations of some key parameters, such as exploitation-vs-exploration trade-off, learning back-up and discount rates, and speed of learning, and interact over a very large number of games on a grid infrastructure; experimental data is then analysed to generate clusters of agents that demonstrate interesting associations between eventual performance ranking and learning parameters’ set-up. The evolution of these clusters indicates that agents with a predisposition to knowledge exploration and slower learning tend to perform better than exploiters, which tend to prefer fast learning. Observing these clusters vis-à-vis the playing behaviours of the agents makes it also possible ...

Realistic Simulation of Cultural Heritage

IGI Global eBooks, 2020

One of the most challenging problems in the simulation of real environments is to generate worlds... more One of the most challenging problems in the simulation of real environments is to generate worlds that appear realistic and more attractive. It becomes increasingly challenging when the simulated environment focuses on minors (students), because the young generation has high demands on simulation systems due to their experience in computer gaming. Virtual museums are among the most important simulation environments, which present cultural and educational content for everyone. Their purpose is to enrich the users experience by allowing an intuitive interaction with the museum artifacts and to offer knowledge with the most pleasant ways. This paper focuses on the aspects of realistic simulations in the development of virtual 3D environments for Cultural Heritage applications. This study includes aspects regarding some of the most high-tech image effects, applicable artificial intelligence methods, powerful game engines, how real object can be reconstructed realistically and how all those features may be combined to produce realistic, pleasant, productive and educative environments.

Enhanced Virtual Reality Experience in Personalised Virtual Museums

IGI Global eBooks, 2020

Thisarticlefocusesonimportantfactorsinthecreationofenhancedpersonalisedexperiencesin virtualenvir... more Thisarticlefocusesonimportantfactorsinthecreationofenhancedpersonalisedexperiencesin virtualenvironmentsforculturalheritageapplications,especiallythosetargetingvirtualmuseumsand exhibitions.Someofthemostimportantfactorsrelatingtopersonalisedvirtualmuseumsthatrelate tointelligentcontentandusermodellinginvirtualenvironmentsarebeinghighlightedanddiscussed. After an extensive review of the current trends in the domain, the article presents a generalised frameworkforthedevelopmentofthenextgenerationenhancedVRexperiencesinpersonalised virtualmuseums.Thisframework,whichnaturallysurfacesfromthedomain,wasputtothetestin thedevelopmentoftheDynaMusplatformandtwocasestudiesbasedonthisplatformarereferenced andcommentedtosupportsuchanapproach.Thisconceptcanserveasthegeneralframeworkfor developingenhancedpersonalisedvirtualenvironmentsforculturalheritageapplications.

Augmenting existing food image datasets with Greek dishes

Zenodo (CERN European Organization for Nuclear Research), Oct 21, 2021

1 Extensive study has been done in the framework of the relevant Greek national project GRE-Taste... more

Collective artificial intelligence and personalized educational interactions in learning new skills

Η Κοινωνική Μάθηση (Social Learning) αποτελεί μια νέα, σημαντική περιοχή έρευνας στον τομέα της Τ... more Η Κοινωνική Μάθηση (Social Learning) αποτελεί μια νέα, σημαντική περιοχή έρευνας στον τομέα της Τεχνητής Νοημοσύνης (Artificial Intelligence) και συγκεκριμένα στην μελέτη των μηχανισμών μάθησης των πολυπρακτορικών συστημάτων (Multi-Agent Systems) παιχνιδιού. Επίσης, θεωρείτε ένας σημαντικός άξονας μελέτης και εφαρμογών για διάφορους επιστημονικούς τομείς, όπως η κοινωνιολογία, η οικονομία και το μάρκετινγκ. Οι κοινωνικοί οργανισμοί (Social Organizations), βασιζόμενοι στα συστήματα προσομοίωσης με πράκτορες, καθώς και η μελέτη/διερεύνηση των στρατηγικών/τεχνικών εκπαίδευσης και μάθησης των πρακτόρων, εμπνεύστηκαν από την ικανότητα των ανθρώπων να μαθαίνουν από κοινωνικά περιβάλλοντα που είναι πλούσια σε κρυφές πληροφορίες και σε αλληλεπιδράσεις μεταξύ οντοτήτων (π.χ. ανθρώπων). Αυτός ο πλούτος του περιβάλλοντος, είναι μια πηγή πολυπλοκότητας όπου ένας μαθητευόμενος πρέπει να μπορεί να πλοηγηθεί ορθά και αποτελεσματικά σύμφωνα με τον στόχο του. Η εισαγωγή της κοινωνικής δυναμικής στα πολυπρακτορικά περιβάλλοντα, θεωρείται ως μια πολλά υποσχόμενη προσέγγιση για την προσομοίωση ρεαλιστικών συμπεριφορών παιχνιδιού, συμπεριφορές που μοιάζουν με ανθρώπινες, καθώς και για την ανάδειξη της σημαντικότητας της Συλλογικής Γνώσης (Collective Knowledge).Η διατριβή αυτή, στοχεύει στην έρευνα της απόδοσης των συνθετικών πρακτόρων (Synthetic Agents) σε σενάρια μάθησης και παιξίματος σε παιχνίδια στρατηγικής με μηδενικό άθροισμα (Zero-Sum). Επίσης, αναδεικνύουμε την σημαντικότητα και την ικανότητα των μοντέλων μάθησης με βάση τον αντίπαλο (Opponent Based Learning), που παρουσιάζουν τα ανταγωνιστικά κοινωνικά περιβάλλοντα με στόχο την αποδοτικότερη μάθηση και παίξιμο ενός πράκτορα.Οι συνθετικοί πράκτορες σχεδιάστηκαν και αναπτύχθηκαν με βάση διάφορους συνδυασμούς κάποιων βασικών παραμέτρων μάθησης, όπως η επιλογή μεταξύ εξερεύνησης και εκμετάλλευσης γνώσης, ταχύτητα μάθησης κ.α.. Στη συνέχεια, οι συνθετικοί πράκτορες αλληλεπιδρούν σε μεγάλους αριθμούς παιχνιδιών και τα πειραματικά δεδομένα αναλύονται για τον προσδιορισμό των ομαδοποιήσεων τους, που παρουσιάζουν σημαντικές συσχετίσεις μεταξύ των παραμέτρων μάθησης και την τελική κατάταξη των επιδόσεων τους.Επιπλέον, εξετάζουμε πώς οι συνθετικοί πράκτορες αλληλεπιδρούν σε κοινωνικά περιβάλλοντα, χρησιμοποιώντας διάφορες στρατηγικές εκπαίδευσης εναντίον διαφόρων αντιπάλων (πράκτορες με διαφορετικά χαρακτηριστικά μάθησης και παιξίματος). Αυτός ο πειραματισμός εκπαίδευσης και παιξίματος, αναδεικνύει πως η ποιότητα του παιξίματος εξαρτάται περισσότερο από τη ορθή παραμετροποίηση του μηχανισμού μάθησης παρά από την εμπειρία.Εξετάζουμε επίσης, την πρόοδο/εξέλιξη της μάθησης των άπειρων πρακτόρων σε κοινωνικά περιβάλλοντα ανταγωνιστικών παιχνιδιών, στοχεύοντας στον προσδιορισμό της επίδρασης ενός έμπειρου αντιπάλου σε έναν αρχάριο πράκτορα. Η διερεύνηση των επιπτώσεων της πολυπλοκότητας του περιβάλλοντος στη συμπεριφορά παιξίματος και μάθησης των συνθετικών πρακτόρων, αναδεικνύει τον τρόπο με τον οποίο ένας αποτελεσματικός παίκτης πρέπει να προσαρμόσει το προφίλ μάθησης και παιξίματος του, για να διατηρήσει ένα συγκεκριμένο προφίλ απόδοσης, όταν η πολυπλοκότητα τους κοινωνικού περιβαλλοντικές μεταβάλλεται.Για την επιβεβαίωση των πειραμάτων αποτελεσμάτων, διεξάγουμε μια αξιολόγηση, μεγάλης κλίμακας, των υπαρχουσών μεθόδων αξιολόγησης της επίδοσης των παικτών (Elo και Glicko), εφαρμοσμένες σε πολυπρακτορικά συστήματα, όπου τονίζεται μια ασυνέπεια (ασυμφωνία) ως προς τον τρόπο με τον οποίο οι προαναφερθέν μέθοδοι αξιολογούν της επιδόσεις των συνθετικών πρακτόρων. Στην συνέχεια, προτείνουμε μία νέα προσέγγιση αξιολόγησης της επίδοσης των πρακτόρων. Η προσέγγιση αυτή, μπορεί να θεωρηθεί ως βάση για την ανάπτυξη μεθόδων αξιολόγησης της επίδοσης παικτών για πολυπρακτορικά συστήματα. Με ένα μεγάλο αριθμό πειραμάτων αποδεικνύεται η αποτελεσματικότητά της προσέγγισης που προτάχθηκε.Επίσης, στα πλαίσια της διατριβής αυτής, παρουσιάζεται μία μέθοδος επιλογής αντιπάλου από μία ομάδα διαφορετικών πρακτόρων, με σκοπό την αποτελεσματικότερη μάθηση και παίξιμο. Μελετήθηκε ένας συνθετικός πράκτορας (με προφίλ καλού παίκτη) όπου πειραματίζεται σε παιχνίδια στρατηγικής με διάφορες αλληλουχίες διαφορετικών αντιπάλων (πράκτορες με διαφορετικά χαρακτηριστικά παιξίματος και επιδόσεις) για τον προσδιορισμό της αλληλουχίας που θα του προσφέρει την &quot;αποτελεσματικότερη μάθησης&quot;. Τα αποτελέσματα των πειραμάτων, δείχνουν ότι η αποτελεσματικότερη πρόοδος και η σταθερότερη εξέλιξη ενός πράκτορα, προκύπτει όταν αυτός αντιμετωπίζει αρχικά αντιπάλους με χειρότερο προφίλ παιξίματος από το δικό του και σταδιακά επιλέγει αντιπάλους με καλύτερο προφίλ παιξίματος. Επίσης, οι μελέτες δείχνουν πως ένας συνθετικός πράκτορας με σωστά διαμορφωμένο/παραμετροποιημένο μηχανισμό μάθησης αποδίδει καλύτερα όταν αντιμετωπίζει λιγότερο ευνοϊκά διαμορφωμένους πράκτορες.Η προσομοίωση των κοινωνικών περιβαλλόντων, όπως στα προαναφερθέντα πειράματα, απαιτεί τεράστιους υπολογιστικούς πόρους. Για τον λόγο αυτό τα πειράματα της…

Deep Learning-Based Black Spot Identification on Greek Road Networks

Data, Jun 16, 2023

This article is an open access article distributed under the terms and conditions of the Creative... more

Fat calculation from raw-beef-steak images through machine learning approaches: an end-to-end pipeline

Proceedings of the 26th Pan-Hellenic Conference on Informatics, Nov 25, 2022

eTHOS: a web-based tax morale system

International Journal of Electronic Governance

In this paper, an innovative web-based cross-platform information system, named eTHOS, is introdu... more In this paper, an innovative web-based cross-platform information system, named eTHOS, is introduced allowing citizens to either report various complaints, in real-time, about tax evasion incidents, or to highlight good tax behavior practices. This is achieved through the creation of interactive maps that highlight geospatial financial factors for the general public. The main purpose of the system is to capture specific information which will be used in order to understand the various tax morale approaches that lead citizens to file a report (complaint or suggestions for reward) based on their interaction with businesses. Furthermore, the new proposed system aims to gain an operational role as the first citizen oriented digital tool against tax evasion. The contribution of eTHOS to the economy could be important, since it could lead to the containment of tax evasion and the decrease of the VAT Gaps.

Semantic Image Segmentation with Deep Learning for Vine Leaf Phenotyping

IFAC-PapersOnLine

Plant phenotyping refers to a quantitative description of the plant's properties, however in imag... more Plant phenotyping refers to a quantitative description of the plant's properties, however in image-based phenotyping analysis, our focus is primarily on the plant's anatomical, ontogenetical and physiological properties. This technique reinforced by the success of Deep Learning in the field of image based analysis is applicable to a wide range of research areas making high-throughput screens of plants possible, reducing the time and effort needed for phenotypic characterization. In this study, we use Deep Learning methods (supervised and unsupervised learning based approaches) to semantically segment grapevine leaves images in order to develop an automated object detection (through segmentation) system for leaf phenotyping which will yield information regarding their structure and function. In these directions we studied several deep learning approaches with promising results as well as we reported some future challenging tasks in the area of precision agriculture. Our work contributes to plant lifecycle monitoring through which dynamic traits such as growth and development can be captured and quantified, targeted intervention and selective application of agrochemicals and grapevine variety identification which are key prerequisites in sustainable agriculture.

Συλλογική τεχνητή νοημοσύνη και προσωποποιημένες εκπαιδευτικές αλληλεπιδράσεις στην εκμάθηση νέων δεξιοτήτων

Social Learning has been an exciting area in Artificial Intelligence for researching learning mec... more Social Learning has been an exciting area in Artificial Intelligence for researching learning mechanisms for in multi-agent game playing systems, besides being a useful concept in diverse science areas, like sociology, economics and marketing. The agent-based simulation of social organizations, based on the investigation of agents’ training and learning tactics and strategies, has been inspired by the ability of humans to learn from social environments which are rich in interactions between peers and partial or hidden information. Such richness is a source of complexity that an effective learner has to be able to navigate. Introducing socially inspired dynamics in multi-agent environments is perceived as a promising approach to simulate realistic human-like playing behaviors and highlight the importance of collectively gained knowledge.In this thesis, we first investigate the performance of synthetic agents in playing and learning scenarios in a turn-based zero-sum game and we highl...

Multi-agents based virtual environments for cultural heritage

2017 XXVI International Conference on Information, Communication and Automation Technologies (ICAT)

This paper focuses on the integration of artificial intelligence methods in multi-agent virtual e... more This paper focuses on the integration of artificial intelligence methods in multi-agent virtual environments for cultural heritage applications and especially virtual museums and exhibitions. Multi-agent systems are considered here as being autonomous social organizations capable of developing dynamic personalized virtual environments. In this respect, some of the most common artificial intelligence methods in intelligent virtual environments are discussed. Three different sub-domains of intelligent virtual environments are identified and presented, aiming at an enhanced virtual reality experience. The paper concludes by highlighting the importance of artificial intelligence applications in cultural heritage, and the dynamics of the virtual museums with personalized content, based on the profile of the user and the need of the case.

Sustainable Ecotourism through Cutting-Edge Technologies

Sustainability, 2022

Tourism is a phenomenon that dates back to ancient times. Ancient Greek philosophers recognised, ... more Tourism is a phenomenon that dates back to ancient times. Ancient Greek philosophers recognised, adopted, and promoted the concept of rest-based tourism. Ecotourism is a particular type of tourism that connects with activities that take place in nature, without harming it, along with the herbal and animal wealth. According to estimates, the global ecotourism industry is currently booming due to various reasons, and it is becoming an important factor of sustainable regional development. This article presents the vision, work, and outcomes of project AdVENt, a project focusing natively in sustainable ecotourism through natural science and technological innovation. AdVENt’s study area includes the National Parks of Oiti (or Oeta) and Parnassus in Central Greece, where there is a remarkable native flora with a high endemism rate integrated with areas of cultural value and national and European hiking routes and paths of varying difficulty.

Deep Learning Platforms in Food Recognition

24th Pan-Hellenic Conference on Informatics, 2020

Image content prediction with novel deep learning approaches is a hot research topic for many sci... more Image content prediction with novel deep learning approaches is a hot research topic for many scientific disciplines. Image-based food types recognition and their ingredients is a particularly challenging task, since food dishes are typically deformable objects, usually including complex semantics, which makes the task of defining their structure very difficult. In this paper we introduce a novel web-based system that exploits commercial deep learning based platforms to predict image content for food recognition. In addition to that, the system combines the individual predictions of the platforms to produce more accurate results. Also, we provide a short assessment of the chosen platforms to determine the most efficient one, in the domain of food recognition. The paper is concluded by highlighting the key features of platforms and the advantages of the new presented system.

Personalized Interactive Edutainment in Extended Reality (XR) Laboratories

ERCIM News, 2020

AnCoins: Image-Based Automated Identification of Ancient Coins Through Transfer Learning Approaches

Pattern Recognition. ICPR International Workshops and Challenges, 2021

Metis: Multi-Agent Based Crisis Simulation System

With the advent of the computational technologies (Graphics Processing Units - GPUs) and Machine ... more With the advent of the computational technologies (Graphics Processing Units - GPUs) and Machine Learning, the research domain of crowd simulation for crisis management has flourished. Along with the new techniques and methodologies that have been proposed all those years, aiming to increase the realism of crowd simulation, several crisis simulation systems/tools have been developed, but most of them focus on special cases without providing users the ability to adapt them based on their needs. Towards these directions, in this paper, we introduce a novel multi-agent-based crisis simulation system for indoor cases. The main advantage of the system is its ease of use feature, focusing on non-expert users (users with little to no programming skills) that can exploit its capabilities a, adapt the entire environment based on their needs (Case studies) and set up building evacuation planning experiments with some of the most popular Reinforcement Learning algorithms. Simply put, the syste...

Crowd simulation for crisis management: The outcomes of the last decade

Machine Learning with Applications, 2020

The last few decades, crowd simulation for crisis management is highlighted as an important topic... more The last few decades, crowd simulation for crisis management is highlighted as an important topic of interest for many scientific fields. As the continues evolution of computational resources increases, along with the capabilities of Artificial Intelligence, the demand for better and more realistic simulation has become more attractive and popular to scientists. Along those years, there have been published hundreds of research articles and have been created numerous different systems that aim to simulate crowd behaviors, crisis cases and emergency evacuation scenarios. For better outcomes, recent research has focused on the separation of the problem of crisis management, to multiple research sub-fields (categories), such as the navigation of the simulated pedestrians, their psychology, the group dynamics etc. There have been extended research works suggesting new methods and techniques for those categories of problems. In this paper, we propose three main research categories, each one consist of several sub-categories, relying on crowd simulation for crisis management aspects and we present the outcomes of the last decade, focusing mostly on works exploiting multi-agent technologies. We analyze a number of technologies, methodologies, techniques, tools and systems introduced throughout the last years. A comparative review and discussion of the proposed categories is presented towards the identification of the most efficient aspects of the proposed categories. A general framework, towards the future crowd simulation for crisis management is presented based on the most efficient to yield the most realistic outcomes of the last decades. The paper is concluded with some highlights and open questions for future directions.

Rating the skill of synthetic agents in competitive multi-agent environments

Knowledge and Information Systems, 2018

A very effective and promising approach to simulate real-life conditions in multi-agent virtual e... more A very effective and promising approach to simulate real-life conditions in multi-agent virtual environments with intelligent agents is to introduce social parameters and dynamics. Introduction of social parameters in such settings reshapes the overall performance of the synthetic agents, so a new challenge of reconsidering the methods to assess agents' evolution emerges. In a number of studies regarding such environments, the rating of the agents is being considered in terms of metrics (or measures or simple grading) designed for humans, such as Elo and Glicko. In this paper, we present a large-scale evaluation of existing rating methods and a proposal for a new rating approach named Relative Skill-Level Estimator (RSLE), which can be regarded as a base for developing rating systems for multi-agent systems. The presented comparative study highlights an inconsistency in rating synthetic agents in the context described by the widely used methods and demonstrates the efficiency of the new RSLE.

On Structure-From-Motion Application Challenges

International Journal of Computational Methods in Heritage Science, 2017

Numerous software solutions that implement the Structure-from-Motion/Multi-View Stereo (SfM/MVS) ... more Numerous software solutions that implement the Structure-from-Motion/Multi-View Stereo (SfM/MVS) 3D reconstruction approach have been made available during the last decade. These allow the production of high quality in terms of geometry and colour information 3D models with the use of unordered image collections that depict a static scene or object from arbitrary viewpoints. Nowadays SfM/MVS-based 3D reconstruction approaches constitute a popular solution in a variety of applications within the cultural heritage domain. As with all 3D reconstruction approaches it has its limitations and application challenges. In this work, the authors attempt to compose a set of guidelines that are based on the important outcomes of important published works that propose solutions that overcome the various challenges introduced by non-friendly surface types.

Using opponent models to train inexperienced synthetic agents in social environments

2016 IEEE Conference on Computational Intelligence and Games (CIG), 2016

This paper investigates the learning progress of inexperienced agents in competitive game playing... more This paper investigates the learning progress of inexperienced agents in competitive game playing social environments. We aim to determine the effect of a knowledgeable opponent on a novice learner. For that purpose, we used synthetic agents whose playing behaviors were developed through diverse reinforcement learning set-ups, such as exploitation-vs-exploration trade-off, learning backup and speed of learning, as opponents, and a self-trained agent. The paper concludes by highlighting the effect of diverse knowledgeable synthetic agents in the learning trajectory of an inexperienced agent in competitive multiagent environments.

Synthetic learning agents in game-playing social environments

Adaptive Behavior, 2016

This paper investigates the performance of synthetic agents in playing and learning scenarios in ... more This paper investigates the performance of synthetic agents in playing and learning scenarios in a turn-based zero-sum game and highlights the ability of opponent-based learning models to demonstrate competitive playing performances in social environments. Synthetic agents are generated based on a variety of combinations of some key parameters, such as exploitation-vs-exploration trade-off, learning back-up and discount rates, and speed of learning, and interact over a very large number of games on a grid infrastructure; experimental data is then analysed to generate clusters of agents that demonstrate interesting associations between eventual performance ranking and learning parameters’ set-up. The evolution of these clusters indicates that agents with a predisposition to knowledge exploration and slower learning tend to perform better than exploiters, which tend to prefer fast learning. Observing these clusters vis-à-vis the playing behaviours of the agents makes it also possible ...

THE 'SYNTHESIS' VIRTUAL MUSEUM

by Mediterranean Archaeology and Archaeometry (MAA) and Chairi Kiourt

Open linked data technologies pave the way towards the semantic Web of the future by a) exploitin... more Open linked data technologies pave the way towards the semantic Web of the future by a) exploiting the abundance in data availability, b) enhancing the continuing application developments in the Web and computer technologies, c) increasing the availability of game engines towards an expansion of techniques and d) bridging culture and education with gaming. In this context, we introduce an innovative and content-dynamic web-based virtual museum, which relies and exploits the rich content of distributed web cultural resources and supports the creation of custom painting virtual exhibitions for cultural and educational purposes based on gaming technologies.

Η αρχαιολογική έρευνα στη Θάσο μέσα από το ψηφιακό αποθετήριο ΑRΕΝΑ (http://arena.ipet.gr )

by Despoina Tsiafakis, Natasa Michailidou, Chairi Kiourt, and Ioannis Mourthos

Πρακτικά 8ου Συμποσίου Θασιακών Μελετών, Η Θάσος διά μέσου των αιώνων: Ιστορία – Τέχνη - Πολιτισμός Θάσος 12-14 Οκτωβρίου 2019, 2020

The paper present aspects of the archaeological research on the island of Thasos through the info... more The paper present aspects of the archaeological research on the island of Thasos through the information and bibliography collected within the repository of the ARENA project (Archaeological REsearch in the North Aegean).
The main theme of ARENA (http://arena.ipet.gr/ ) is the collection, management and presentation of the archaeological research conducted in Aegean Thrace and generally the northern Greek territory. Its goal is the transformation of the traditional research into digital archaeology accessible and usable by everyone.

CSEDU 2024 AI-Enabled Art Education: Unleashing Creative Potential and Exploring Co-Creation Frontiers

by Vassilis Evangelidis and Chairi Kiourt

16th International Conference on Computer Supported Education Angers - France May 2 - 4,, 2024

Abstract: This paper explores the intersection of AI technology and art education with the goal o... more Abstract: This paper explores the intersection of AI technology
and art education with the goal of unleashing creative potential
and delving into co-creation frontiers within culturally diverse
societies. Emphasizing the significance of arts in fostering social
inclusion and embracing cultural diversity, the paper outlines the
transformative role of AI in enhancing creativity and collaborative
processes within education. The envisioned ecosystem, integrating AI-assisted co-creation tools, storytelling development, and
digital display boards, proposes a comprehensive framework for
revolutionizing art education. The paper places a strong emphasis
on ethical considerations and technical robustness, delineating
key challenges and foundational components for successful
implementation. Envisioned as a versatile platform designed for
diverse educational settings, from primary to higher education
and non-formal contexts, the initiative aims to foster cultural
diversity, social inclusion, and tolerance. Simultaneously, it seeks
to promote lifelong learning and collaboration on a global scale.