Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)
Permanent URI for this collection
Browsing Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ) by Title
Now showing 1 - 20 of 144
Results Per Page
- ItemOpen AccessAlgorithms and techniques for efficient data management in the Web
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Νοδαράκης, Νικόλαος; Τσακαλίδης, Αθανάσιος; Τσακαλίδης, Αθανάσιος; Σιούτας, Σπύρος; Τζήμας, Γιάννης; Γαροφαλάκης, Ιωάννης; Χατζηλυγερούδης, Ιωάννης; Μακρής, Χρήστος; Μποζάνης, Παναγιώτης; Nodarakis, NikolaosThe term of cloud computing refers to the usage of computational resources (on software and hardware level) that consist a unified service over a network, like internet. Cloud computing becomes more and more popular among data management and storage applications, because of its ability of handling extremely large amount of data (TB or even PB). Daily, new problems arise that require efficient and scalable solutions for monitoring, processing and storing big data volumes. The most popular and notably efficient tools are key-value stores, that allow unstructured data storage, and large-scale distributed processing systems, like MapReduce. In the context of this thesis, we focus on the proposing techniques that deal with computationally intensive problems. Many centralized approaches have been developed for these problems, but when the data size grows exponentially these algorithms stop being effective. They either fail to confront the problem or need an excessive amount of time to fulfill their goal. It is more than clear that there is an imperative need to turn to distributed and high-scalable solutions that run on a cluster of computers.
- ItemOpen AccessAllergic rhinitis passive detection, assessment, and monitoring using artificial intelligence of things (AIoT) and crowdsensing
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Τζαμαλής, Παντελής; Tzamalis, PantelisAllergy-related respiratory diseases, such as asthma and rhinitis that coexist with the term respiratory allergy, are a major and growing public health problem in Greece and worldwide. Today, 70 million Europeans suffer from chronic asthma and 100 million from allergic rhinitis. Among them, a significant percentage suffer from a severe form of allergic disease, which affects their productivity and quality of life. These numbers are expected to increase in the coming decades, establishing respiratory allergy as a pandemic. It is estimated that the cost of rhinitis in Europe amounts to more than 100 billion euros per year. The main goals of the respective treatment are summarized in the monitoring and control of the symptoms, their treatment, and the effort to prevent future seizures. At the same time, the causal association with allergens is sought, in order to follow a special desensitization treatment in them. However, monitoring patients with allergic respiratory diseases is notably difficult. The aim of this work is to create an integrated, sensor- and crowdsourcing-based, eHealth/mHealth approach that is deployed in wireless environments, which adopts the principles of Healthcare 5.0 for the intelligent, automatic, holistic, effective, and continuous monitoring of allergic rhinitis disease outbreaks. In this paradigm shift into the digital healthcare approach, a study is conducted which is distinguished into two main parts. Initially, a platform is delivered that is capable of the large-scale spatiotemporal detection and monitoring of the disease exacerbation, in real-time. Additionally, the association of the allergens' onset with the allergic disease symptoms and the levels of occurrence of various irritants in a region (humidity, dust) are examined. The data that feeds the platform is generated from multiple and different types of resources in the form of, sensor measurements that are distributed across a region, text data from social media, as well as data that comes from the users’ subjective assessments, combined with geolocation recordings, concerning the intensity of their allergic symptoms. An analysis mechanism is integrated into the platform, which can process the hybrid forms of data that include, sensor measurements analysis, text-mining, and subjective inputs analysis, where the latter is a part of the users’ participatory sensing. Afterward, by the developed visualization mechanisms, the platform’s role is to provide information, in an easy-to-understand way, about the patient's health status and statistical inferences in charts and plots regarding the various forms of allergic symptoms and the people that affect them. Additionally, a notification service is deployed in case of intense symptoms. It is obvious that this approach integrates both human and machine intelligence together with their hybrid interaction. Nevertheless, the real challenge is to provide a complete automated analysis and passive monitoring of the disease, based only on machine intelligence. For that purpose, the work that the second part of the study consists of is the design and development of another platform that aims at the intelligent and automatic evaluation and identification of kinesiological data that is related to an individual’s allergic rhinitis symptoms (such as scratching the nose). Thus, this part of the study is transferred from the large-scale to the individual-scale disease monitoring scenario. In particular, the motion data, in this study, is actually gestures that are retrieved from smart wrist-worn devices. A whole set of algorithmic components is developed for an end-to-end analysis in this phase. In particular, an innovative data processing pipeline is employed in association with the utilization of AI. Specifically, the usage of both statistical learning models and cutting-edge neural network architectures leads to the practical motion data evaluation and pattern recognition of allergic gestures in the users’ daily activities. As a case study, the introduced end-to-end machine learning pipeline is integrated and tested for its efficiency, for the first time, in a real-world scenario, in the context of the development of a national funded project, called Personal Allergy Tracer, where the multidisciplinarity is adopted by collaborating with recognized allergists that validated the whole approach in real patients via a pilot phase. Additionally, as this thesis is a part of the project, another system is deployed, which is related to the individual’s allergic rhinitis status non-invasive monitoring, by utilizing all the data resources that the project owns. In particular, except for the motion data that pertains to the development of this thesis and is obtained by the smart wearables, the rest of the data that is exploited by this system, is retrieved by a smartphone application that is a major component of the project, and corresponds to: a) the analysis of voice alteration of the user that is obtained through the smartphones’ microphone, and, b) the subjective evaluation by the users regarding the intensity of their symptoms. Access to such data takes place through collaboration with established partners in the IT industry, in Greece. This multi-source data is then analyzed in a hybrid manner, and finally, the system induces the automated monitoring of respiratory allergy and acts as a sentinel and disease prevention tool for patients with allergic rhinitis symptoms. However, because AI is a major component in the various tasks of the analysis of this thesis, a framework has been developed capable of handling multi-domain end-to-end machine learning pipelines regarding the motion data evaluation and pattern recognition, as well as the classification of text data that is related to social media posts. The framework automates all the cutting-edge procedures from data processing, model training, fine-tuning, evaluation, and validation of the whole pipeline in the domains of time-series analysis and text-mining, and provides a Prediction Service for automatically deploying the pipeline in production. In conclusion, various benefits can arise from such an analysis of both approaches. For instance, the complementary collected information, from the crowdsourced data which constitutes the individuals’ subjective self-assessment and social media posts, as well as the sensors’ measurements, can lead to the better control and management of seasonal symptoms in cases of allergic diseases, where a medical decision support system can be formulated. The automated, passive, even geolocated, recording of symptoms’ exacerbations in combination with automated notification services can contribute significantly to the control of the disease, reducing morbidity and improving the quality of life of patients with respiratory allergy and their performance. It also has a positive impact on maintaining the productive capacity of patients with respiratory allergies at work or school.
- ItemOpen AccessCooperative and cognitive communication techniques for wireless networks
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)(2014-05-16) Τσίνος, Χρήστος; Μπερμπερίδης, Κωνσταντίνος; Μπερμπερίδης, Κωνσταντίνος; Θεοδωρίδης, Σέργιος; Ροντογιάννης, Αθανάσιος; Βαρβαρίγος, Εμμανουήλ; Κωτσόπουλος, Σταύρος; Τουμπακάρης, Δημήτριος-Αλέξανδρος; Κρικίδης, Ιωάννης; Tsinos, ChristosDuring the past years wireless communications have been exhibiting an increased growth rendering them the most common way for communication. The continuously increasing demand for wireless services resulted in limited availability of the wireless spectrum. To this end, Cognitive Radio (CR) techniques have been proposed in literature during the past years. The concept of CR approach is to utilize advanced radio and signal-processing technology along with novel spectrum allocation policies to enable new unlicensed wireless users to operate in the existing occupied spectrum areas without degrading the performance of the existing licensed ones. Moreover, the broadcast and fading nature of the wireless channel results in severe degradation on the performance of wireless transmissions. A solution to the problem is the use of multiple-antenna systems so as to achieve spatial diversity. However, in many cases, the communication devices' nature permit the support of multiple antennas due to size, power consumption, and hardware limitations. To this end, cooperative communications provide an alternative way to achieve spatial diversity via virtual antenna arrays formed by single antenna nodes. It is noteworthy that cooperation has an important role within the CR literature as many techniques developed within its context exploiting the benefits of cooperation in order to achieve improved performance. Therefore, the aim of the present dissertation is to develop efficient and practical cognitive, cooperative and cognitive cooperative schemes. More specifically the contributions are the following ones. The first contribution is a novel CR communication scheme. During the past years numerous CR communication schemes have been presented in literature. To the best of our knowledge, the majority of them were developed assuming perfect Channel State Information (CSI) at the unlicensed user's side. There are several cases where the licensed users do not desire any interaction with the unlicensed ones. In such cases, the assumption that the unlicensed user can obtain CSI that concerns the licensed user channels is not valid and as a result the corresponding communication technique cannot be applied. Therefore, at first we propose an novel CR communication scheme that requires CSI that can be estimated in a completely blind manner. Then, the corresponding blind estimation scheme is developed. Another significant contribution is the theoretical results that have been derived for both the perfect CSI case and the imperfect CSI case (when the blind estimation scheme is employed for obtaining the corresponding CSI). Especially, the theoretical results that concern the imperfect CSI case are some of the first ones that appear in the relevant literature, to the best of our knowledge. The second contribution is a decentralized adaptive Eigenvalue-Based spectrum Sensing (EBSS) technique for multi-antenna CR Systems. Spectrum Sensing is a fundamental functionality in CR systems. In general, the unlicensed user employs a spectrum sensing technique in order to detect licensed user(s) activity in scenarios where the former user is permitted to establish a communication link only via spectrum areas that are temporarily free of the latter one's transmissions. EBSS techniques are known to achieve good performance and also to be applicable in a completely blind manner. In the literature so far, only batch and centralized cooperative EBSS techniques have been considered which, however, suffer from limitations that render them impractical in several cases such as, when time-varying channels are involved or continuous spectrum monitoring is required. Thus, the aim here is to develop practical cooperative adaptive versions of typical Eigenvalue-Based Spectrum Sensing (EBSS) techniques which could be applied in a completely decentralized manner and cope well in time-varying scenarios. To this end, at first, novel adaptive EBSS techniques are developed for the Maximum Eigenvalue Detector (MED), the Maximum-Minimum Eigenvalue Detector (MMED), and the Generalized Likelihood Ratio Test (GLRT) schemes, respectively, for a single-user (no cooperation) case. Then, a novel distributed subspace tracking method is proposed which enables the cooperating nodes to track the joint subspace of their received signals. Based on this method, cooperative decentralized versions of the adaptive EBSS techniques are subsequently developed that overcome the limitations of the existing batch centralized approaches. The third contribution is a new cooperative scheme for half-duplex uplink transmission. The technique is based on a virtual MIMO structure formed by the single antenna source and relays nodes along with the multi-antenna base station which is the destination node. The new technique aims at providing increased diversity and multiplexing gains, contrariwise to existing approaches where the proposed techniques achieve increased diversity gain at the cost of severe multiplexing gain loss. The theoretical outage probability and the corresponding Diversity Multiplexing Trade-off (DMT) curve of the proposed technique are also derived. The final contribution is two novel algorithms which enable the relay cooperation for the distributed computation of the beamforming weights in a blind and adaptive manner, without the need to forward the data to a fusion center. The proposed algorithms are constituent parts of two corresponding distributed beamforming schemes for relay networks that distribute the computational overhead equally among the relay nodes. In the first scheme, the beamforming vector is computed through minimization of the total transmit power subject to a receiver quality-of-service constraint (QoS). In the second scheme, the beamforming weights are obtained through maximization of the receiver signal-to-noise-ration (SNR) subject to a total transmit power constraint. The proposed approaches achieve close performance to the one of the optimal beamforming solutions derived assuming perfect channel state information at the relays' side.
- ItemOpen AccessData mining system for tree and network structures in medical images
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)(2014-11-24) Σκούρα, Αγγελική; Μεγαλοοικονόμου, Βασίλειος; Τσακαλίδης, Αθανάσιος; Νικηφορίδης, Γεώργιος; Χατζηλυγερούδης, Ιωάννης; Κωσταρίδου, Ελένη; Δερματάς, Ευάγγελος; Bakic, Predgrad; Skoura, AggelikiΑνατομικές δομές με δενδρική τοπολογία απαντώνται συχνά στο ανθρώπινο σώμα και οπτικοποιούνται σε ιατρικές εικόνες χρησιμοποιώντας απεικονιστικές τεχνικές με ακτίνες-χ και τη χρήση σκιαγραφικού υλικού. Χαρακτηριστικά παραδείγματα τέτοιων δομών είναι το βρογχικό δένδρο εντός των πνευμόνων το οποίο οπτικοποιείται με εικόνες αξονικής τομογραφίας και τα γαλακτοφόρα δένδρα εσωτερικά του μαστού τα οποία οπτικοποιούνται με γαλακτογραφίες. Σκοπός της παρούσας διδακτορικής διατριβής αποτελεί η ανάπτυξη ενός συνόλου αλγοριθμικών μεθόδων για την αυτοματοποίηση της ανάλυσης των ανατομικών δομών του ανθρωπίνου σώματος που έχουν τοπολογία δένδρου ή τοπολογία δικτύου. Πιο συγκεκριμένα, οι δύο βασικοί στόχοι της διατριβής είναι η ανάπτυξη μεθόδων ειδικά σχεδιασμένων για τη ψηφιακή επεξεργασία των ιατρικών εικόνων που απεικονίζουν δομές με διακλαδώσεις και η ανάπτυξη μεθοδολογικών πλαισίων για τη διερεύνηση της σχέσης μεταξύ τοπολογίας και παθοφυσιολογίας αυτού του τύπου ανατομικών δομών. Το πρώτο κεφάλαιο της διατριβής παρουσιάζει μια βιβλιογραφική ανασκόπηση σχετικά με τις ανατομικές δομές του ανθρωπίνου σώματος με τοπολογία διακλαδώσεων καθώς και το κίνητρο για την παρούσα έρευνα. Οι επιμέρους ερευνητικοί στόχοι, οι κύριες συνεισφορές και η γενικότερη απήχηση της διατριβής αναφέρονται επίσης. Το δεύτερο κεφάλαιο εστιάζει στην κατάτμηση εικόνας. Η κατάτμηση εικόνας αποτελεί το πρώτο βήμα στη διαδικασία ανάλυσης ιατρικών εικόνων και στα συστήματα αναγνώρισης προτύπων και οι αλγόριθμοι κατάτμησης αποτελούν κρίσιμα τμήματα των σύγχρονων ιατρικών διαγνωστικών συστημάτων. Παρά την πλούσια βιβλιογραφία στην περιοχή, η ανάγκη για αποδοτικές μεθοδολογίες κατάτμησης εφαρμόσιμες σε μεγάλο εύρος απεικονιστικών τεχνικών παραμένει. Προσπαθώντας να αντιμετωπιστεί αυτή η ερευνητική πρόκληση, μια καινοτόμα και πλήρως αυτοματοποιημένη μεθοδολογία για την κατάτμηση των δενδρικών ανατομικών δομών παρουσιάζεται. Η βασική ιδέα είναι ο συνδυασμός τεχνικών ανίχνευσης ακμών με μεθόδους ανάπτυξης περιοχών για να επιτευχθεί αποδοτική κατάτμηση. Η υβριδική αυτή προσέγγιση εφαρμόστηκε και αξιολογήθηκε σε δύο σύνολα δεδομένων ιατρικών εικόνων από διαφορετικές απεικονιστικές τεχνικές (γαλακτογραφίες και αγγειογραφίες) και η απόδοσή της συγκρίθηκε με τεχνικές κατάτμησης της υπάρχουσας τεχνολογικής στάθμης. Το τρίτο κεφάλαιο επικεντρώνεται στην ανίχνευση των κόμβων διακλάδωσης το οποίο συνιστά ένα σημαντικό υπολογιστικό στάδιο στα πλαίσια της επεξεργασίας των ιατρικών εικόνων που απεικονίζουν δομές δενδρικής τοπολογίας. Οι κόμβοι διακλάδωσης αποτελούν σημεία-κλειδιά για τον προσδιορισμό της θέσης του δένδρου και η σωστή ανίχνευσή τους είναι ένα σημαντική για την αυτοματοποίηση διαδικασιών επεξεργασίας εικόνας όπως ευθυγράμμιση εικόνας, κατάτμηση εικόνας και ανάλυση των προτύπων διακλάδωσης. Ωστόσο, η ανάπτυξη αυτοματοποιημένων τεχνικών για την ανίχνευση των κόμβων διακλάδωσης δυσχεραίνεται από τα διαφορετικά επίπεδα θορύβου που υπάρχουν κατά μήκος της δενδρικής δομής. Η προτεινόμενη μεθοδολογία ανίχνευσης απαρτίζεται από δύο κύρια στάδια: ανίχνευση γωνιακών σημείων σε διάφορες κλίμακες και προσδιορισμό της θέσης της διακλάδωσης. Η βασική συνεισφορά της νέας μεθοδολογίας είναι η χρήση ενός τοπικά προσαρμοζόμενου κατωφλιού κατά τη φάση της ανίχνευσης προκειμένου να αντιμετωπιστεί αποδοτικά η ανίχνευση των σημείων διακλάδωσης που βρίσκονται στα χαμηλά δενδρικά επίπεδα. Η αξιολόγηση της μεθόδου πραγματοποιήθηκε χρησιμοποιώντας ένα σύνολο δεδομένων από κλινικές γαλακτογραφίες και η απόδοσης της συγκρίνεται με αντίστοιχες τεχνικές της υπάρχουσας τεχνολογικής στάθμης. Στο τέταρτο κεφάλαιο παρουσιάζονται καινοτόμες μεθοδολογίες για τον χαρακτηρισμό και την κατηγοριοποίηση των ανατομικών δενδρικών δομών στοχεύοντας στη διερεύνηση της συσχέτισης μεταξύ τοπολογίας και παθολογίας των αντίστοιχων οργάνων. Οι μέθοδοι περιλαμβάνουν κατηγοριοποίηση χρησιμοποιώντας περιγραφικά χαρακτηριστικά της τοπολογίας όπως η δενδρική ασυμμετρία, η χωρική κατανομή των σημείων διακλάδωσης, η στρεβλότητα των κλάδων και άλλα γεωμετρικά χαρακτηριστικά του δένδρου. Επιπρόσθετα σε αυτό το κεφάλαιο, ένα νέο μεθοδολογικό πλαίσιο προτείνεται για την ανάλυση δενδρικών τοπολογιών χρησιμοποιώντας διανύσματα που κωδικοποιούν τις σχέσεις παιδιού-γονέα των κόμβων και ελαστικό ταίριασμα μεταξύ των ακολουθιών. Η υπεροχή της νέας αυτής μεθόδου έναντι των μεθόδων της υπάρχουσας τεχνολογικής στάθμης για την κατηγοριοποίηση δένδρων αξιολογήθηκε πειραματικά ως προς ευαισθησία, ειδικότητα και ακρίβεια. Στο πέμπτο κεφάλαιο μελετώνται τεχνικές συλλογικής μάθησης. Η ενοποίηση πολλαπλών αλγορίθμων μηχανικής μάθησης συνιστά σημαντική πρόοδο για τις μεθοδολογίες κατηγοριοποίησης και βασίζεται στην ιδέα του συνδυασμού των προβλέψεων ενός πλήθους κατηγοριοποιητών με σκοπό τη μεγιστοποίηση της ακρίβειας κατηγοριοποίησης. Τρεις τεχνικές συνδυαστικής μάθησης βασισμένες στην τεχνική της ενδυνάμωσης (boosting) και η χρήση ενός συνδυαστικού κανόνα που ονομάζεται Πρότυπο Απόφασης (Decision Template) χρησιμοποιούνται για τη βελτιστοποίηση της ακρίβειας που επιτυγχάνουν οι κατηγοριοποιητές βάσης. Τα πειραματικά αποτελέσματα επιβεβαιώνουν την υπεροχή των μεθόδων συλλογικής μάθησης. Κλείνοντας, τα συμπεράσματα της διατριβής παρουσιάζονται στο έκτο κεφάλαιο. Οι περιορισμοί των προτεινόμενων τεχνικών καθώς και οι προοπτικές για επιπρόσθετη ερευνητική εργασία αναλύονται.
- ItemOpen AccessDesign and analysis of algorithms for non-cooperative environments
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Βουδούρης, Αλέξανδρος Ανδρέας; Καραγιάννης, Ιωάννης; Καραγιάννης, Ιωάννης; Κακλαμάνης, Χρήστος; Νικολετσέας, Σωτήριος; Ζαρολιάγκης, Χρήστος; Κοσμαδάκης, Σταύρος; Μαρκάκης, Ευάγγελος; Σπυράκης, Παύλος; Voudouris, Alexandros A.This thesis studies issues related to problems that arise in large-scale distributed environments with non-cooperative users, who act strategically and compete with each other to maximize their personal payoffs. For instance, imagine a scenario where a set of users compete over a resource, such as the bandwidth of a communication link or advertisement slots when keywords are queried in search engines on the Internet. A mechanism takes input from all participating users (which represents their preferences) and outputs an allocation of the resource to them (it distributes the bandwidth or assigns slots). Each user aims to select her input to the mechanism in order to satisfy her personal objectives (possibly by misreporting her true preferences), without caring about the social welfare which we would like to maximize as the designers of the mechanism. Therefore, this behavior induces a strategic game among the users who act as players and sequentially change their strategies until they reach an equilibrium state (if one exists) from which no one has any incentive to deviate. Due to the strategic behavior of the users, the equilibrium that is reached may be of low quality in terms of some objective function like the social welfare, compared to what could happen if a central authority dictated the strategies of the users. The price of anarchy and stability are two quantification measures of this kind of inefficiency at equilibrium. Our main goal in this thesis is to understand the advantages and constraints of the strategic games that arise in non-cooperative environments as means of computation. What can they compute and how well can they compute it? Is it possible to alter the rules of the game and incentivize the players to truthfully report their preferences? We answer to such questions related to equilibrium computation, price of anarchy and stability estimation, and truthful mechanism design for many interesting and important classes of problems. In particular, we study resource allocation with budget constraints, opinion formation, ownership transfer with expert advice, and revenue maximization in randomized combinatorial sales.
- ItemOpen AccessDesign and evaluation of wireless energy transfer protocols in ad hoc networks
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Ράπτης, Θεοφάνης; Νικολετσέας, Σωτήριος; Νικολετσέας, Σωτήριος; Μπερμπερίδης, Κωνσταντίνος; Βαρβαρίγος, Εμμανουήλ; Γαροφαλάκης, Ιωάννης; Κακλαμάνης, Χρήστος; Τσακαλίδης, Αθανάσιος; Βλάχος, Κυριάκος; Raptis, TheofanisIn this dissertation we investigate problems which arise from the application of wireless energy transfer in ad hoc and sensor networks and we address those problems by designing efficient protocols. The dissertation is comprised of 10 Chapters. Chapters 1 and 2 provide an introduction to the subject and to the relevant state of the art. In Chapters 3-9 the reader can find the scientific core of the dissertation. Insights for future work are provided in Chapter 10. Finally, there are also two Appendices, one with a list of scientific publications which were produced during the writing of this work and another with an extended abstract of the dissertation in Greek language. The scientific core of the dissertation is divided in two parts. Chapters 3-6 focus on mobile wireless charging in stationary sensor networks, while Chapters 7-9 investigate different models and alternative applications of wireless energy transfer in ad hoc networks. More specifically, at the first part, in Chapter 3, we focus on wireless charging with a single Mobile Charger assuming uniform network deployments and in Chapter 4 on non-uniform network deployments. Wireless charging with multiple Mobile Chargers is presented in Chapter 5 and further investigated in Chapter 6 by introducing hierarchical collaboration. The different models, which are introduced in the second part of the dissertation, consist of interactive wireless charging in populations of mobile peers in Chapter 7, radiation aware wireless charging in Chapter 8 and a realistic model which takes into account fundamental properties of superposition of electromagnetic fields in Chapter 9.
- ItemOpen AccessDesign, development and evaluation on decision making algorithms, based on innovative smart energy networks
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Μαμουνάκης, Ιωάννης; Γαροφαλάκης, Ιωάννης; Γαλλόπουλος, Ευστράτιος; Βαρβαρίγος, Εμμανουήλ; Μεγαλοοικονόμου, Βασίλειος; Μακρής, Χρήστος; Μπερμπερίδης, Κωνσταντίνος; Ζαρολιάγκης, Χρήστος; Mamounakis, IoannisThe use of renewable energy sources across Europe and many other parts of the world has been largely supported by government and economic incentives in recent years. Taking into account European Union climate change legislation requiring EU Member States to produce 20% of their electricity from renewable sources by the year 2020, it is clear that few European countries will be able to achieve this objective. The only way in which the majority of EU countries can get closer to the goal is to make use of excess capacity from domestic Renewable Energy Sources (RES). This thesis examined the need to involve small and medium-sized producers and consumers in the energy market with the aim both of economic benefit and of changing their energy behavior. Initially, the current techniques / algorithms were studied regarding the grouping of producers and consumers as well as the pricing process. It was proposed to create a new smart grid architecture with the contribution of a new aggregator that organizes all small and medium-sized producer-consumers in clusters that act on the same policy. Studies have been made using already known algorithms (spectral, genetic, and adaptive) to find the appropriate producer-consumer cluster that will reduce their energy costs. The study was enriched with the use of prediction techniques, offering even greater accuracy in the end result. Clustering methods based on correlation were then proposed to find the appropriate producer / consumer groups to serve current demand for both real-time and post-day markets. The pricing of the energy supply on the market was an important part of this dissertation, for which purpose the following algorithms were proposed. Initially, a pricing algorithm was developed and implemented in flexible markets for the economic benefits and smooth operation of the market. An algorithm of producer-consumer grouping in smart power networks was then studied and implemented with the aim of reducing consumption and changing energy behavior. All of the above-mentioned technical algorithms were evaluated using experiments from anonymous user data over the last five years. The integration of the algorithms into two platforms VIMSEN Decision Support System (VIMSEN-DSS) platform and the Researchers Algorithm Toolkit (RAT) platform has demonstrated both the correct operation of the outputs and the smooth operation of the proposed techniques within an integrated information system.
- ItemOpen AccessDistributed processing techniques for parameter estimation and efficient data-gathering in wireless communication and sensor networks
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)(2015-05-07) Bogdanovic, Nikola; Μπερμπερίδης, Κωνσταντίνος; Νικολετσέας, Σωτήριος; Ροντογιάννης, Αθανάσιος; Θεοδωρίδης, Σέργιος; Βαρβαρίγος, Εμμανουήλ; Ψαράκης, Εμμανουήλ; Κοφίδης, Ελευθέριος; Μπογκντάνοβιτς, ΝίκολαThis dissertation deals with the distributed processing techniques for parameter estimation and efficient data-gathering in wireless communication and sensor networks. With the aim of enabling an energy aware and low-complexity distributed implementation of the estimation task, several useful optimization techniques that generally yield linear estimators were derived in the literature. Up to now, most of the works considered that the nodes are interested in estimating the same vector of global parameters. This scenario can be viewed as a special case of a more general problem where the nodes of the network have overlapped but different estimation interests. Motivated by this fact, this dissertation states a new Node-Specific Parameter Estimation (NSPE) formulation where the nodes are interested in estimating parameters of local, common and/or global interest. We consider a setting where the NSPE interests are partially overlapping, while the non-overlapping parts can be arbitrarily different. This setting can model several applications, e.g., cooperative spectrum sensing in cognitive radio networks, power system state estimation in smart grids etc. Unsurprisingly, the effectiveness of any distributed adaptive implementation is dependent on the ways cooperation is established at the network level, as well as the processing strategies considered at the node level. At the network level, this dissertation is concerned with the incremental and diffusion cooperation schemes in the NSPE settings. Under the incremental mode, each node communicates with only one neighbor, and the data are processed in a cyclic manner throughout the network at each time instant. On the other hand, in the diffusion mode at each time step each node of the network cooperates with a set of neighboring nodes. Based on Least-Mean Squares (LMS) and Recursive Least-Squares (RLS) learning rules employed at the node level, we derive novel distributed estimation algorithms that undertake distinct but coupled optimization processes in order to obtain adaptive solutions of the considered NSPE setting. The detailed analyses of the mean convergence and the steady-state mean-square performance have been provided. Finally, different performance gains have been illustrated in the context of cooperative spectrum sensing in cognitive radio networks. Another fundamental problem that has been considered in this dissertation is the data-gathering problem, sometimes also named as the sensor reachback, that arises in Wireless Sensor Networks (WSN). In particular, the problem is related to the transmission of the acquired observations to a data-collecting node, often termed to as sink node, which has increased processing capabilities and more available power as compared to the other nodes. Here, we focus on WSNs deployed for structural health monitoring. In general, there are several difficulties in the sensor reachback problem arising in such a network. Firstly, the amount of data generated by the sensor nodes may be immense, due to the fact that structural monitoring applications need to transfer relatively large amounts of dynamic response measurement data. Furthermore, the assumption that all sensors have direct, line-of-sight link to the sink does not hold in the case of these structures. To reduce the amount of data required to be transmitted to the sink node, the correlation among measurements of neighboring nodes can be exploited. A possible approach to exploit spatial data correlation is Distributed Source Coding (DSC). A DSC technique may achieve lossless compression of multiple correlated sensor outputs without establishing any communication links between the nodes. Other approaches employ lossy techniques by taking advantage of the temporal correlations in the data and/or suitable stochastic modeling of the underlying processes. In this dissertation, we present a channel-aware lossless extension of sequential decoding based on cooperation between the nodes. Next, we also present a cooperative communication protocol based on adaptive spatio-temporal prediction. As a more practical approach, it allows a lossy reconstruction of transmitted data, while offering considerable energy savings in terms of transmissions toward the sink.
- ItemOpen AccessEffective algorithms and improved high-volume data analysis techniques
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Λιαπάκης, Ξενοφών; Παυλίδης, Γεώργιος; Παυλίδης, Γεώργιος; Μεγαλοοικονόμου, Βασίλης; Γαροφαλάκης, Ιωάννης; Τζήμας, Ιωάννης; Σιούτας, Σπύρος; Τσόλης, Δημήτρης; Στυλιάρας, Γεώργιος; Liapakis, XenofonBehind the buzzword big data lays a challenge which is very true and very important both for academia and industry. Efficiently handling an overwhelming volume of data, possibly coming from a large number of sources abiding to vastly different functionality protocols, coding conventions, sampling rates, and quality standards, is very important from engineering, algorithmic, economic, and even social perspective. However, no matter how challenging the technical part is, it is nonetheless only a fraction of the entire challenge. This happens because harnessing new, non-trivial knowledge from literally an entire data ocean is even more challenging given the additional constraint that the value of this newly obtained knowledge must at least equal the total extraction it, including factors such as power, storage cost, equipment procurements, and data collection cost. And this is the marginal case, which cannot be maintained indefinitely. On the contrary, the knowledge value must be a multiple of the total effort cost in order for any big data pipeline to be viable from a business perspective. The twofold objective of this PhD dissertation is to: To explore the applications of parallelism to accelerating critical computations in challenging problems from various fields. One very concrete example comes from the emerging field of computational combinatorics. Specifically, a novel graph structural resilience metric based on triangles and paths is proposed. Since this metric is purely structural, namely function oblivious, it can be applied to virtually any graph as long as the patterns it relies on have a physical meaning. To show how parallelism can be part of very efficient and wide applicable computational kernels, such as those found in the BLAS library for basic linear algebra operations, which can be applied to various engineering and financial problems, the proposed algorithms are examined from a computational kernel perspective. It is shown that they can be applied to other problems as well, increasing thus their usefulness.
- ItemOpen AccessEfficient algorithms for big data management
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Δρίτσας, Ηλίας; Dritsas, EliasIn the context of the doctoral research, I dealt with data management problems by developing methods and techniques that, on the one hand, maintain or improve the privacy and anonymity of users and, on the other hand, are efficient in terms of time and storage space for large volumes of databases. The research results of the work focus on the following: Evaluate the performance of queries in a large volume database using or not the Bloom Filter structure. Evaluate workload time, memory and disk usage of the Privacy Preserving Record Linkage (PPRL) problem in Hadoop MapReduce Framework. Methods of answering queries of nearest neighbors to spatio-temporal data (moving users trajectories) in order to preserve anonymity, where queries are applied to clustered or non-clustered data. The k anonymity method was used, where, the set of anonymity with which each moving object of the space-time database is being camouflaged, consists of its k nearest neighbors. The robustness of the method was quantified with a probability of 1/k and the effect of dimensionality and correlation of the data on the preservation of anonymity and privacy was studied. The above method was improved in terms of efficient storage of spatio-temporal data by applying queries of nearest neighbors to Hough transformed nonlinear trajectories of moving objects. The application of secure k-NN queries was evaluated in the GeoSpark environment. Sentiment Analysis on Twitter Data and Tourist Forecasting at Apache Spark
- ItemOpen AccessEfficient data management for the Internet of Things
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Αμαξηλάτης, Δημήτριος; Ζαρολιάγκης, Χρήστος; Ζαρολιάγκης, Χρήστος; Amaxilatis, DimitriosTechnological developments in recent years in both hardware and software have led to an explosion of devices and services in what is often called the Internet of Things. Today in every home in the US there are 11 "smart" devices with sensors capable of accurately describing and watching their environment. This trend is expected to continue further in the future, with this figure rising to 500 in 2022, sharply increasing the volume of information that can be exported from human activity. This plethora of data creates enormous potential for developing applications and services for users to improve their life parameters (from personal comfort to health or transport). The secure and effective processing of these data is a major problem, for which no specific solution has been widely adopted so far to address the significant problems of personal data protection and the efficient management of such data. Technologies such as IPv6 promote cloud management by directly exposing all devices over the Internet and thus allowing easy interaction and remote control, as opposed to Bluetooth LE, where local device management is the only available solution and selected information is published on the Internet using gateway devices. Finally, it is particularly important to enrich the data with external data such as meteorological forecasts or data on traffic and public transport routes. In this thesis, we are studying the real-time monitoring and organization of sensing infrastructures according to the changing requirements of applications and users. The aim is to design, implement and evaluate experimental self-organizing mechanisms using semantic information to improve the quality of data flows provided at the Internet level. As part of this process, we also seek to combine flows to more efficiently share and manage information. The mechanisms we implement are based on the "semantic entities" model, either as a part of the device network or as part of a web service aiming at balancing computing and storage requirements at the various levels of the network hierarchy. The goal is also to study new data processing and tracking techniques to draw appropriate conclusions, predictions and decisions. In this thesis, we studied Internet of Things sensing infrastructures of multiple sizes in terms of the technologies and methods that are needed to appropriately organize and analyze data the data streams that are form from the installed devices to the cloud level. The aim was to design, implement and evaluate by experimental methods flexible mechanisms for representing the semantic information of the installations, generating statistical analysis on the data flows and finally provide a better understanding on the conditions of the environments monitored. As part of this process, the analysis of the data flows is used to generate “knowledge” and find problematic and defective devices or calculate predictions on their future values.
- ItemOpen AccessEfficient mining of design patterns for advanced modeling and optimization of content management system-based web applications
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Γκαντούνα, Βασιλική; Σιούτας, Σπυρίδων; Σιούτας, Σπυρίδων; Γαροφαλάκης, Ιωάννης; Τζήμας, Ιωάννης; Χατζηλυγερούδης, Ιωάννης; Λυκοθανάσης, Σπυρίδων; Μακρής, Χρήστος; Παυλίδης, Γεώργιος; Gkantouna, VassilikiContent Management Systems (CMSs) play an increasingly important role in the evolution of the World Wide Web, since almost half of the websites today use some form of CMS as their main development platform. CMSs provide development teams with standardized software platforms that significantly facilitate and speed up Web development, while maintaining high quality and low-cost implementation without requiring extensive programming expertise. Due to their flexibility and ease of use, CMSs are nowadays constantly gaining in popularity, and undoubtedly, they are among the most prominent platforms used by developers for building Web applications. The goal of this dissertation is to study and address challenging problems in the area of the widely used category of Web applications developed on top of CMSs, the so-called CMS-based Web applications, contributing novel models, methodologies and tools to optimize the quality of both Web development process and final products. In the first part of this thesis, we deal with issues concerning the model-driven engineering perspective in CMS development. Despite the widespread use of CMSs, current Model-Driven Web Engineering (MDWE) methodologies have ignored this phenomenon, and thus, they cannot support the automated (model-driven) generation of CMS-based Web applications. Given that MDWE methodologies are actually driven by the expressiveness of the modeling languages that are being used within their context, this failure of existing MDWE methodologies is probably caused by the fact that they are based on generic modeling languages which lack the expressiveness to capture the particular development context of CMS platforms. To address this problem, we propose a new genre of modeling languages, called CMS-oriented modeling languages, which are specifically defined over the particular development context of CMS platforms. More specifically, we provide a framework to support their definition in three main stages. The first stage involves the analysis of the target CMS platform under a number of different viewpoints in order to determine its particular development context. Then, the second stage involves the creation of the domain model for the target CMS, including all the key elements of its development context that need to be captured by the CMS-oriented modeling language. Finally, the third involves the definition of the CMS-oriented modeling language, i.e., the formalization of its modeling primitives and concepts, based on the specifications elaborated in the two previous stages. In this way, the proposed CMS-oriented modeling languages have the expressiveness to represent CMS-based Web applications as models consisting of modeling elements which are in direct correspondence with the elements specified in the actual development context of the target CMS, making the transition from modeling to source-code level considerably easier, enabling the model-driven development in CMS domain. To support our case, we have applied the proposed framework on the popular open-source Joomla! CMS platform resulting in the definition of a modeling language oriented towards Joomla! CMS, which is accompanied by a prototype tool for modeling Joomla!-based Web applications according to the specifications of the proposed language. In the second part, we deal with issues concerning the design quality of CMS-based Web applications throughout their entire development lifecycle. Undoubtedly, design quality is a key determinant factor for the success of an application, since it has a direct impact on usability, performance, maintenance, and other quality attributes of the application. However, as CMS-based Web applications evolve during their lifecycle, their design quality is typically severely undermined by several factors, such as their constantly increasing intrinsic complexity and the continuous modifications due to changing requirements, to name a few. In fact, as the lifecycle of a CMS-based Web application evolves, design inconsistencies tend to occur more and more frequently and it is very common for the final product to often drift away from the original design. This has a significant negative impact on usability and overall application quality, implying the need for techniques and tools to support the development of CMS-based Web applications by ensuring that design quality is maintained throughout all the phases of the application lifecycle, from the early design phases up to development and maintenance. In response to this need, we propose a novel methodology for the automated design quality evaluation of CMS-based Web applications, which, by inspecting their conceptual model under the viewpoint of design reuse, automatically identifies potential problems in them that compromise design quality by means of efficiency, consistency, and usability. More specifically, the proposed methodology automatically extracts the conceptual model of a CMS-based web application and, subsequently, submits it to a pattern-based analysis in order to identify the occurrences of all the incorporated recurrent patterns implying design reuse, appropriate or not. Then, based on evaluation metrics, the identified patterns are categorized as effective or ineffective design solutions. Intuitively, this categorization results in the identification of “hot-spots”, i.e., problematic areas in the application model due to ineffective ad-hoc forms of reuse, as well as, of a set of effective reusable design structures that can be used as building blocks by designers for the refactoring of such areas. By applying the methodology on a CMS-based Web application, developers can gain important information regarding its design quality, since it automatically highlights to them potential problems in the design model and provides them with a set of refactoring recommendations for improving the application’s structure, consistency and usability. In the third and last part, we deal with issues concerning the identification of domain-specific design patterns for the development of CMS-based Web applications. Domain-specific design patterns are a powerful tool for developing quality Web applications in a certain application domain, since they provide designers with proven solutions to recurring design problems that occur in the domain. By using them, developers can gain a number of important advantages, such as increased design quality, reduced development and maintenance costs, as well as better communication between interdisciplinary teams, since patterns provide a common vocabulary to discuss the various design alternatives. Nevertheless, despite their numerous advantages, the number of available domain-specific patterns today is still small, as well as, the number of the application domains that have been explored which is also very limited. The most important factor contributing to this is the absence of systematic techniques to assist domain designers with the pattern identification process. This is mainly due to the fact that this process cannot be easily carried out in an automated way, since it is hard to find a systematic way to detect well-defined design partitions of a domain, able to capture its semantics. Furthermore, it is the very nature of domain-specific patterns that makes their identification difficult, since they must encompass generality and variability for being able to be instantiated in various web applications in a certain application domain. As a result, there is a great need for techniques to support the identification of domain-specific web design patterns. To this end, we here report the first attempt to automatically support their identification, by providing designers with recommendations about candidate domain-specific design patterns. More specifically, we propose a methodology which captures the designs of a collection of websites in a target application domain, and then, by utilizing graph mining techniques, obtains the identification of all the reusable design solutions which are used in them by designers as building blocks for addressing typical domain problems. The latter are provided to designers as candidate recommendations and they can significantly facilitate the identification of domain-specific web design patterns. By inspecting them, domain designers can explore all the various design alternatives which are used in the domain for handling a certain recurring problem. This way, they can have an overview of all the common design practices used in the domain for addressing typical domain problems, among which they can possibly recognize best practices and domain-specific design patterns. To support our case, we present the results of a case study conducted for the domain of academic websites.
- ItemOpen AccessEfficient tranceiver [sic] techniques for interference and fading mitigation in wireless communication systems
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Βλάχος, Ευάγγελος; Μπερμπερίδης, Κωνσταντίνος; Ψαράκης, Εμμανουήλ; Ροντογιάννης, Αθανάσιος; Βαρβαρίγος, Εμμανουήλ; Γαλλόπουλος, Ευστράτιος; Αντωνακόπουλος, Θεόδωρος; Κρικίδης, Ιωάννης; Vlachos, EvangelosWireless communication systems require advanced techniques at the transmitter and at the receiver that improve the performance in hostile radio environments. The received signal is significantly distorted due to the dynamic nature of the wireless channel caused by multipath fading and Doppler spread. In order to mitigate the negative impact of the channel to the received signal quality, techniques as equalization and diversity are usually employed in the system design. During the transmission, the phenomenon of inter-symbol interference (ISI) occurs at the receiver due to the time dispersion of the involved channels. Hence, several delayed replicas of previous symbols interfere with the current symbol. Equalization is usually employed in order to combat the effect of the ISI. Several implementations for equalization filters have been proposed, including linear and non-linear processing, providing complexity-performance trade-offs. It is known that the length of the equalization filter determines the complexity of the technique. Since the wireless channels are characterized by long and sparse impulse responses, the conventional equalizers require high computational complexity due to the large size of their filters. In this dissertation, we have further investigated the long standing problem of equalization in light of the recently derived theory of compressed sampling (CS) for sparse and redundant representations. The developed heuristic algorithms for equalization, can exploit either the sparsity of the channel impulse response (CIR), or the sparsity of the equalizer filters, in order to derive efficient implementation designs. To this end, building on basis pursuit and matching pursuit techniques new equalization schemes have been proposed that exhibit considerable computational savings, increased performance properties and short training sequence requirements. Our main contribution for this part is the Stochastic Gradient Pursuit algorithm for sparse adaptive equalization. An alternative approach to combat ISI is based on the orthogonal frequency division multiplexing (OFDM) system. In this system, the entire channel is divided into many narrow subchannels, so as the transmitted signals to be orthogonal to each other, despite their spectral overlap. However, in the case of doubly selective channels, the Doppler effect destroys the orthogonality between subcarriers. Thus, similarly to ISI, the effect of intercarrier interference (ICI) is introduced at the receiver, where symbols which belong to other subcarriers interfere with the current one. Considering this problem, we have developed iterative algorithms which recursively cancels the ICI at the receiver, providing performance-complexity trade-offs. For low or medium Doppler spreads, the typical approach is to approximate the frequency-domain channel matrix with a banded one. On this premise, we derived reduced-rank preconditioned conjugate gradient (PCG) algorithms in order to estimate the equalization matrix with a reduced number of iterations. Also developed an improved PCG algorithm with the same complexity order, using the Galerkin projections theory. However, in rapidly changing environments, a severe ICI is introduced and the banded approximation results in significant performance degradation. In order to recover this performance loss, we developed regularized estimation framework for ICI equalization, with linear complexity with respect the the number of the subcarriers. Moreover, we proposed a new equalization technique which has the potential to completely cancel the ICI. This approach works in a successive manner through a number of stages, conveying from the fully-connected ordered successive interference cancellation architecture (OSIC) in order to fully suppress the residual interference at each stage of the equalizer. On the other hand, diversity can improve the performance of the communication system by sending the information symbols through multiple signal paths, each of which fades independently. One approach to obtain diversity is through cooperative transmission, considering a group of nearby terminals (relays) as forming one virtual antenna array and applying a spatial beamforming technique so as to optimize the communication via them. Such beamforming techniques differ from their classical counterparts where the array elements are located in a common processing unit, due to the distribution of the relays in the space. In this setting, we developed new distributed algorithms which enable the relay cooperation for the computation of the beamforming weights leveraging the computational abilities of the relays. Each relay can estimate only the corresponding entry of the principal eigenvector, combining data from its network neighbours. The proposed algorithms are applied to two distributed beamforming schemes for relay networks. In the first scheme, the beamforming vector is computed through minimization of a total transmit power subject to the receiver quality-of-service (QoS) constraint. In the second scheme, the beamforming weights are obtained through maximization of the receiver SNR subject to a total transmit power constraint. Moreover, the proposed algorithms operate blindly, implying that no training data are required to be transmitted to the relays, and adaptively, exhibiting a quite short convergence period.
- ItemOpen AccessEfficient wireless power transfer protocols in rechargeable sensor networks
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Μάδια, Αντελίνα; Νικολετσέας, Σωτήριος; Βαρβαρίγος, Εμμανουήλ; Μπερμπερίδης, Κωνσταντίνος; Βλάχος, Κυριάκος; Κακλαμάνης, Χρήστος; Καραγιάννης, Ιωάννης; Κουλουρίδης, Σταύρος; Madhja, AdelinaIn this thesis, we study problems related to the efficient energy management of sensor networks and propose protocols that exploit the Wireless Power Transfer (WPT) technology using magnetic resonant coupling. A common, realistic assumption of all problems that we investigate is that the amount of available energy supplies is finite; hence, it is crucial to manage it in the most efficient way in order to prolong the network lifetime, and achieve other required network properties as well. The problems on which we focus in this thesis can be classified in two main categories. In the first one, special network entities (called chargers) are required to store large amount of energy supplies and appropriately transfer it to the network devices. In the second category, we study different network models where no special devices are used for the energy management; in contrast, the ordinary network devices are capable to charge each other in a peer-to-peer manner. In particular, we first study wireless rechargeable sensor networks using multiple mobile chargers. In such a setting, the most crucial questions that we aim to answer are the following two. How can the mobile chargers coordinate in order to partition the network area amongst them in fair way, based on the energy supplies they have? Further, what trajectory should each of them follow in order to traverse its assigned subregion and charge the sensor nodes therein? To answer these questions, we propose both centralized and distributed coordination solutions that exploit different levels of knowledge. Moreover, we propose a hierarchical collaborative charging scheme where the chargers are partitioned in two types, the mobile chargers, which are responsible for replenishing the energy of the sensor nodes, and the special chargers, which are responsible for recharging the mobile chargers. In this setting, we investigate how the special chargers must coordinate with each other, which trajectories they should follow, and how much energy must be transferred to each mobile charger. We also focus on the problem of selecting the appropriate power of a single static charger over time in mobile ad hoc networks in order to adapt to the mobility and energy consumption characteristics of the mobile nodes, as these are revealed in an online manner. Finally, we study the problem of designing interaction protocols for networks that consist of computationally weak devices, which are able to exchange energy in a peer-to-peer manner, thus eliminating the need for special entities, like the chargers. The objective is to let the devices distributively construct specific network structures (such as star and tree networks) and achieve appropriate energy distributions. For each of these problems we propose and analyze the properties (both theoretically and experimentally) of various protocols, which utilize different levels of knowledge about the network.
- ItemOpen AccessInnovative algorithmic techniques in cloud computing
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Πισπιρίγκος, Γεώργιος; Pispirigkos, GeorgiosIn recent years, due to the universal spread of the internet, the amount of digitally available information continuously grows. Indicatively, in 2021 there are 7.83 billion people, 4.66 billions of whom are considered active Internet users 1 . In addition, over 100 million Gigabytes of data have yet been indexed, while approximately 2.5 billion Gigabytes of raw data are generated on daily basis 2 . Thus, as digital information massively grows, the need for compact data representations has never been more immediate. Among the rest data representations, information networks can undoubtedly be considered one of the most prominent, since they manage to harmonically combine different aspects of information in the same data entity. Specifically, this representation model subtly combines different data modalities, so as to conveniently incorporate the intrinsic inference and semantics. Therefore, by presenting each individual entity as a node and by denoting any kind of association with an interconnection edge, information graphs may adeptly formulate any functional system as one of interacting entities. The usage of this general-purpose abstraction has practically been proven beneficial and has widely been adopted in various scientific sectors - such as chemistry, biology, neuroscience, finance, linguistics, social sciences, software engineering, digital marketing, etc., - mostly because of its obvious interpretation. As a result, with the extended application of this hierarchical data representation in various important issues, - such as customer segmentation, epidemiology, political influence evolution, criminal identification, tissue/organ detection, etc.; - graph analysis has attracted significant scientific interest, triggering the conduction of an impressive amount of research. Focusing on social media, where in 2021 more than 4.14 billion people are considered active social media users 1 , it is obvious that the respective information can effectively be structured with information networks. Specifically, each user can compactly be presented as a vertex, while any kind of user interaction - such as “friendship”, “liking”, “following”, “sharing”, “expressing interest”, “re-tweeting”, etc.; - might be expressed with an interconnection edge between the corresponding vertices. From sentiment analysis, expert identification and mood analysis to recommendation systems, digital footprint analysis and viral marketing campaigns, graph theory lies under numerous substantial issues, aiming to introduce efficient information managing and data mining techniques. Undeniably, one of the most important network analysis topic is community detection. This problem principally aims to identify groups of highly similar entities, aka communities, by primarily leveraging the network topology [1-9] of the data graph under study. The necessity of this topic, as plainly explained in , concerns the deeper understanding of the subjacent network structure that could lead to the extraction of advantageous insights regarding the underlying dynamic processes. Therefore, with its appropriate generalization and its widespread adoption from numerous fields - such as recommendation systems, targeted market analysis, influence propagation, link prediction, opinion mining etc.; - this NP-hard graph analysis problem  has unquestionably become one of the most essential and challenging. In respect of sociology, community detection can alternatively be construed as the expression of the homophily effect [11,12], which reflect the natural human tendency to mostly associate and interact with groups of similar. Concentrating on social media, community detection can be differently interpreted as the identification of social media users’ groups that are, either directly or indirectly, connected with each other and tend to interact more often, comparing to users of different communities . Consequently, by identifying the strongly related groups of individuals, any social network might be naturally decomposed to groups of highly interacting entities, the communities, which plainly disclose the given social graph's inner mechanics. Despite the lack of a broadly accepted definition [1-9], a community can be intuitively perceived considering its graph representation. Specifically, a set of vertices is expected to form a meaningful community, if only its intra-cluster and inter-cluster densities are bigger, and smaller respectively, to the average link density of the original graph . However, as it is strongly underlined in , the optimal community hierarchy is generated when the inter-cluster and intra-cluster densities are comparatively better than expected, and not by the optimization of the individuals. In other words, a good graph division would not be the one in which the number of connections between communities is minimized, but the one in which there are fewer inter-connection edges than originally found. Because of this abstraction, community detection has practically become one of the most conceptually challenging and computationally demanding network analysis topics. Due to its profitable application in plenteous scientific areas, a numerous set of community detection algorithms [1-9] has already been published. From methods that are exclusively based on the repetitive calculation of a global network topology criterion, to alternatives inspired by discrete mathematics and physics, the pluralism of classic community detection approaches is indeed remarkable. The great majority of those methods and algorithms are originally designed to be generally applied to any information network. The classic community detection techniques - e.g., the Louvain  algorithm, the Girvan–Newman  algorithm, the Clauset–Newman–Moore  algorithm, the edge centrality optimization method , the geodesic edge betweenness  approach, etc.; - are basically recursive methods of high polynomial computational complexity aiming to optimize an iteratively modified and repetitively calculated set of global network topology metrics, in order to extract the underlying community hierarchy from any possible network. Nevertheless, those methods’ capacity has practically confirmed [1,4,8] to be profoundly limited in terms of scalability, outcome consistency, and overall reliability. As meticulously described in [1-9,17], the classic community detection solutions are principally static methods that neglect not only the information networks’ topological heterogeneities, but also their significantly variant subjacent community structure. Additionally, apart from the undisputed computational difficulties of global optimization processes, the actual size of today’s real-world social networks also acts as another major inhibitory factor. As plainly explained in , with their corresponding computational complexity being at least quadratic, the application of classic community detection algorithms in large-scale information networks, such as contemporary social media graphs, is prohibitive. In this regard, the primary intention of this thesis is the introduction of highly scalable and accurate community detection methodologies that would efficiently extract the underlying community hierarchy of modern, large-scale social information networks regardless of their size and density. Initially, by alternatively interpreting the node-oriented definition of community to its equivalent edge-oriented representation, the introduction of distributed community prediction takes place. The true purpose of these methodologies is to efficiently identify the subjacent community hierarchy of any large-scale social graph through prediction. This is achieved by classifying the imminent graph's edges to either the ones associating nodes of different communities, aka inter-connection edges, or of the same, aka intra-connection edges, solely based on plain network topology characteristics. The promising perspectives of the distributed community prediction have meticulously been analysed in the [18-20] research publications. All different proposed methodologies have thoroughly been examined on numerous real-life social networks and proven superior to various classic community detection methods in terms of performance, stability and accuracy. Furthermore, aiming to enhance the community identification process to further consider different aspects of social networks’ information, such as the intrinsic user profile information, the article  has been published. In this article, a distributed, hybrid, community detection methodology has been introduced that ably combined the local topological characteristics of each social media graph under study along with the existing user profile information, in order to unveil the subjacent community structure. The proposed hybrid community detection approach  has been extensively tested on various real-world social graphs, roundly compared to other classic divisive community detection algorithms and practically proven highly scalable and adequately accurate. Finally, the research publications [22,23] have plainly presented the beneficial application of community detection in other sectors such as the viral marketing and computational linguistics. Particularly, in case of , the Twitter Personality based Communicative Communities Extraction (T-PCCE) system has been introduced. This system , given a real-world Twitter social subgraph, aims to identify the communities of high internal information flows, by also considering the users’ personality traits. Regarding the case of , a word-sense disambiguation methodology has been introduced. Specifically, by employing classic community detection techniques and establishing the unprecedented concept of community coherence on the Wikipedia Entities’ semantic ontologies graph, this distributed methodology demonstrated impressive precision and general computational performance as regards the Wikification / text annotation problem.  Santo Fortunato: Community detection in graphs. CoRR abs/0906.0612 (2009)  Schaeffer, Satu. (2007). Graph Clustering. Computer Science Review. 1. 27-64. 10.1016/j.cosrev.2007.05.001.  Devi, J. & Eswaran, Poovammal. (2016). An Analysis of Overlapping Community Detection Algorithms in Social Networks. Procedia Computer Science. 89. 349-358. 10.1016/j.procs.2016.06.082.  Stavros Souravlas, Angelo Sifaleras, M. Tsintogianni, Stefanos Katsavounis: A classification of community detection methods in social networks: a survey. Int.J. Gen. Syst. 50(1): 63-91 (2021), doi: 10.1080/03081079.2020.1863394  Cai, Q., Ma, L., Gong, M. and Tian, D. (2016) A survey on network community detection based on evolutionary computation, Int. J. Bio-Inspired Computation, Vol. 8, No. 2, pp.84–98, DOI: 10.1504/IJBIC.2016.076329.  Bisma S. Khan, Muaz A. Niazi: Network Community Detection: A Review and Visual Survey. CoRR abs/1708.00977 (2017)  Michele Coscia, Fosca Giannotti, Dino Pedreschi: A Classification for Community Discovery Methods in Complex Networks. CoRR abs/1206.3552 (2012)  Papadopoulos, S., Kompatsiaris, Y., Vakali, A. et al. Community detection in Social Media. Data Min Knowl Disc 24, 515–554 (2012). https://doi.org/10.1007/s10618-011-0224-z  Javed, M.A., Younis, M.S., Latif, S., Qadir, J., Baig, A., Community detection in networks: A multidisciplinary review, Journal of Network and Computer Applications (2018), doi: 10.1016/j.jnca.2018.02.011.  Andrea Lancichinetti, Mikko Kivelä, Jari Saramäki, Santo Fortunato: Characterizing the community structure of complex networks. CoRR abs/1005.4376 (2010)  Khediri, Nourhene & Karoui, Wafa. (2017). Community Detection in Social Network with Node Attributes Based on Formal Concept Analysis, 2017 IEEE/ACS 14th International Conference on Computer Systems and Applications, 1346-1353. 10.1109/AICCSA.2017.200.  Yuan Li: Community Detection with Node Attributes and its Generalization. CoRR abs/1604.03601 (2016)  P. Held, B. Krause and R. Kruse, "Dynamic Clustering in Social Networks Using Louvain and Infomap Method," 2016 Third European Network Intelligence Conference (ENIC), 2016, pp. 61-68, doi: 10.1109/ENIC.2016.017.  Newman MEJ. Modularity and community structure in networks. PNAS June 6, 2006 103 (23) 8577-8582; https://doi.org/10.1073/pnas.0601602103  Vincent D. Blondel, Jean-Loup Guillaume, Renaud Lambiotte, Etienne Lefebvre: Fast unfolding of community hierarchies in large networks. CoRR abs/0803.0476 (2008)  E.J. Newman, Mark & Girvan, Michelle. (2004). Finding and Evaluating Community Structure in Networks. Physical review. E, Statistical, nonlinear, and soft matter physics. 69. 026113. 10.1103/PhysRevE.69.026113.  Peel, L.; Larremore, D.B.; Clauset, A. The ground truth about metadata and community detection in networks. Sci. Adv. 2017, 3, e1602548, doi:10.1126/sciadv.1602548.  Makris, C.; Pettas, D.; Pispirigos, G. Distributed Community Prediction for Social Graphs Based on Louvain Algorithm. In IFIP International Conference on Artificial Intelligence Applications and Innovations; Springer: Cham, Switzerland, 2019; pp. 500–511. DOI: 10.1007/978-3-030-19823-7_42  Makris, C.; Pispirigos, G.; Rizos, I.O. A Distributed Bagging Ensemble Methodology for Community Prediction in Social Networks. Information 2020, 11, 199. https://doi.org/10.3390/info11040199  Makris, C.; Pispirigos, G. Stacked Community Prediction: A Distributed Stacking-Based Community Extraction Methodology for Large Scale Social Networks. Big Data Cogn. Comput. 2021, 5, 14. https://doi.org/10.3390/bdcc5010014  Konstantinos Georgiou, Christos Makris, Georgios Pispirigos: A Distributed Hybrid Community Detection Methodology for Social Networks. Algorithms 12(8): 175 (2019). https://doi.org/10.3390/a12080175  Eleanna Kafeza, Andreas Kanavos, Christos Makris, Georgios Pispirigos, Pantelis Vikatos: T-PCCE: Twitter Personality based Communicative Communities Extraction System for Big Data. IEEE Trans. Knowl. Data Eng. 32(8): 1625-1638 (2020). DOI: 10.1109/TKDE.2019.2906197  Makris, C.; Pispirigos, G.; Simos, M.A. Text Semantic Annotation: A Distributed Methodology Based on Community Coherence. Algorithms 2020, 13, 160. https://doi.org/10.3390/a13070160
- ItemOpen AccessInnovative data mining techniques and applications in social networks
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Βικάτος, Παντελεήμων; Μακρής, Χρήστος; Γαροφαλάκης, Ιωάννης; Μεγαλοοικονόμου, Βασίλειος; Νικολετσέας, Σωτήριος; Χατζηλυγερούδης, Ιωάννης; Καφέζα, Ελεάννα; Σγάρμπας, Κυριάκος; Vikatos, PanteleimonNowadays, people constantly use online social networking sites sharing content about their daily lives and things that happen around them. These systems have revolutionized the way we communicate, by organizing our offline social relationships in a digital form. Simultaneously, the social media serve the intentions of marketers to promote their brands and products due to this massive participation in these platforms and the endless potentials of improving their strategies for more effective marketing campaigns. The selection of the targets, the diffusion of brand promotion messages and the place of advertising content in popular web pages constitute some of most significant objectives of marketers in order to gain the interest of potential customers. Our research objective is to deal marketing campaign tasks using data mining techniques. We explore new methodologies in marketing campaign targeting, conversational and personalized advertising as well as the information propagation in Online Social Networks. We introduce a methodology for calculating user influence and select targets of marketing campaigns using bridge participation in the evolving social graph. We analyze the improvement of social bots infiltration using automated communication skills and we introduce the conversational social bots as advertising content promoters that improve brand engagement. We provide a novel methodology for personalized advertising using hotlink assignment. Our method enhances browsing experience and leads users to certain advertising content through hotlinks. We also introduce a novel methodology to achieve information diffusion within a social graph that activates a realistic number of users. Our approach combines the predicted patterns of diffusion for each node with propagation heuristics in order to achieve an effective cover of the graph. Our methodologies are useful to recommendation systems as well as to marketers who are interested to use data mining techniques to run effective marketing campaigns.
- ItemOpen AccessIoT systems and wireless power transfer protocols in ad-hoc communication networks
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Κατσιδήμας, Ιωάννης; Katsidimas, IoannisThe techniques of energy maintaining and replenishment in wireless communications networks have become particularly popular in recent years. Rapid technological advances in the field of Wireless Power Transmission (Wireless Power Transmission) have a major impact on sensor networks and more generally on energy-restricted communications networks, paving the way for new methods of energy management in wireless systems. Until recently, existing research mainly focused on maximizing network lifetime, improving charging efficiency, minimizing latency during charging, and so on. Most recent research has already begun to examine algorithmic solutions to address the problems that arise. At the same time, both the evolution and the application of ICT systems in very important areas of our lives (industry, smart home, smart cities, etc.) is rapid. IoT devices (devices connected to the internet) must be able to be deployed in any place and access them from everywhere. A large number of these devices perform monitoring and controlling tasks in "smart" applications as well as in difficult-to-access areas. For the successful implementation of these applications, an IoT device should be small and autonomous, while incorporating sensors, data processing and wireless communication capabilities. These simple conditions imply storage and power management limitations for the IoT devices in order to ensure their continued operation, since neither replacing the power cord nor the battery are viable choices under these conditions or simply because of the flexibility (fast installation without cable, no maintenance). Major growth opportunities in the coming years are expected in wireless power transmission and solar harvesting technologies approved for powering IoT devices. This increase is justified by significant advances in material science, engineering and extensive prototyping. Innovative energy storage solutions for IoT have already appeared in the market. In terms of energy management, a range of integrated solutions are available on the market to supply IoT devices that vary from home, up to industrial applications. RF charging, alongside other WPT technologies and ambient energy harvesters are today viable integrated energy sources for small, portable or not, devices. In this thesis, methods of efficient wireless electromagnetic radiation charging have been developed to improve the charging quality of unstructured communication networks and to develop IoT systems that will be able to utilize this technology. In particular, we present novel algorithmic methods based on a new, more realistic and accurate wireless power transfer model, which can efficiently capture constructive and destructive interference of the electromagnetic waves. These methods deal with problems such as i) charger’s power level, phase configuration and deployment towards power maximization, and ii) find low EM radiation paths in WPT systems. Besides the algorithmic perspective, modern energy provisioning approaches aim to combine different technologies that decouple harvesting from the environment and efficiently manage the available energy. Thus, we present novel IoT energy management platforms that integrate both RF-charging and ambient energy harvesting to power sensing and communication devices. Those platforms are utilised in a real application in the context of Smart Roads and are responsible to power a list of sensors and the corresponding communication module in a sufficient way.
- ItemOpen AccessNovel methods for post-manufacturing and in-field testing of VLSI circuits/systems
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Σισμάνογλου, Παναγιώτης; Νικολός, Δημήτριος; Αραπογιάννη, Αγγελική; Βέργος, Χαρίδημος; Καβουσιανός, Χρυσοβαλάντης; Κάγκαρης, Δημήτρης; Τσιατούχας, Γιώργος; Καλλίγερος, Εμμανουήλ; Νικολός, Δημήτρης; Sismanoglou, PanagiotisIn this work, at first we analyze and evaluate, the already known test data compression schemes for post-manufacturing testing and the DFT mechanisms for the enhancement of in-field testing and circuit reliability. We propose a new dictionary based test data compression method for post-manufacturing testing and we discuss several aspect of the testing procedures, as the reusability of the on-chip decoder for the testing of multiple cores, the reusability of the tester vector memory (local memory of the test head), the on-chip implementation cost of the decoder and the test power consumption. Moreover, we improve each of the aforementioned parameters without degradation of the compression and test application time reduction, efficiency of the proposed dictionary coding scheme. We extend our discussion to the in-field testing, revealing that the already known techniques are characterized from their vast intrusiveness to the normal operation of the circuit, resulting to severe performance degradation and, therefore, they are rarely applied (on startup/shutdown of the system and/or on maintenance time windows). To this end, we proposed two new schemes, in the first scheme, we developed a preemptive Built-In Self-Test (BIST) mechanism that ensures none or minimum intrusiveness to the normal operation of the circuit and in the second scheme, we designed a new storage cell for the hardening of the circuit against radiation-induced soft-errors that outperforms the already known designs. Combining the above two schemes, the frequency of in-field testing can be increased for the detection of permanent delay and/or stuck-at faults due to aging, but also the normal operation of the circuit is protected against soft-errors. Therefore, the reliability of the circuit operation is significantly improved.
- ItemOpen AccessOpen learner modeling techniques in a digital educational game and study of motivation mechanisms
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Λεονάρδου, Αγγελική; Leonardou, AngelikiThe student-centered learning approach known as Digital Game-based Learning (DGBL) uses digital games with educational materials to provoke student interest and to enhance their learning effectiveness. Within the two main DGBL parameters, i.e. learning (education) and gaming (fun, entertainment) the Multiplication Game (MG) was developed. MG is an adaptive digital game for developing and practising multiplication competences, which addresses primary school pupils and aims to develop the necessary multiplication skills in a fun and interactive way, in a safe environment with discreet error management. Through the Open Learner Model (OLM) and Open Social Learner Model (OSLM) techniques, metacognitive skills (e.g. self-assessment, self-organization) are promoted, while the opportunity of comparing pupil performance is offered. At the same time, it is possible for this tool to be used by teachers for easy evaluation and detailed monitoring of pupil performance. The MG supports all the participants involved in the learning process: pupil - teacher - classmates, so that it can be easily incorporated in the classroom. According to the research conducted, it was proved that teachers who have a positive attitude and acceptance for digital games, also have a positive view of MG. The provision of OLM data was also shown to enhance motivation, while gamification motivators were positively correlated with pupils' intrinsic motivation.
- ItemOpen AccessPrivacy-aware human presence and movement analysis in controlled and uncontrolled environments
Τμήμα Μηχανικών Η/Υ και Πληροφορικής (ΔΔ)Ιωαννίδης, Δημοσθένης; Λυκοθανάσης, Σπυρίδων; Κάτσικας, Σωκράτης; Τζοβάρας, Δημήτριος; Ζαρολιάγκης, Χρήστος; Παυλίδης, Γεώργιος; Μουστάκας, Κωνσταντίνος; Βότης, Κωνσταντίνος; Ioannidis, DemosthenisIt is undoubtable that if machines could automatically interpret the human presence and movement in our everyday life, many tasks would be transposed and revolutionized. Multiple human tracking (MHT), encompassing human presence and movement analysis, are key components in many computer vision applications, in which the fundamental goal is to automatically segment, capture, recognize and analyze human activities in both controlled and uncontrolled environments. One of the core topics the thesis deals with refers to the research and development of a robust human detection and tracking framework, which exploits the use of privacy-preserving information for supporting emerging application scenarios in the Energy Efficiency domain such as occupancy-driven facility management and the utilization of calibrated occupancy models for building performance simulation and prediction, which are tools increasingly used by the architecture, engineers and construction community at design, refurbishment and operational phases of a building lifecycle. Furthermore, as of today, an increasing number of perimeter security and surveillance firms are now delivering multiple-sensor threat identification and classification toolkits that can recognize unwarranted and suspicious activity in fenced areas. The trend is to introduce cost-effective solutions coupled with privacy-preserving multi-sensorial networks, to further reduce high false alarm rates as well as to research and develop techniques that can interpret activities of moving objects in the perimeter and robust identify suspicious activity, taking into account minimal false alarm rates and operators burden. Within this thesis, a novel framework for the automatic recognition of suspicious activities in outdoor infrared-based video surveillance is presented, in which its discriminating power in classifying high risk activities like trespassing from zero risk activities such as loitering outside the perimeter is outlined. To cope with human presence and movement in outdoor environments, an extended version of the multiple-human tracking framework introduced for controlled environments (i.e. buildings) is employed as an important step in analyzing human behavior in non-controlled environments, fully exploiting privacy-preserving information from infrared video sequences. During this thesis, an extensive experimental evaluation of the proposed methodologies for human detection, tracking and movement analysis in indoor and outdoor environments has been carried out, which effectively illustrated the robustness of the proposed algorithms and their applicability in a range of emerging application scenarios related to human behavior analysis in buildings, energy efficiency in buildings and critical infrastructure perimeter protection. The thesis is concluded with an in-depth discussion summarizing the major achievements of the current work, as well as identifying open issues, limitations and research challenges that could be addressed in future works.