528 research outputs found
The Δρομοδείχτης της Ελλάδος of 1824 and Athanasios Stageirites (Τίτλος περίληψης)
No full textσ. [281]-290Κείμενο στα ελληνικά με περίληψη στα αγγλικά με τον τίτλο: The Δρομοδείχτης της Ελλάδος of 1824 and Athanasios StageiritesThe article first examines the close relationship between the publication “Δρομοδείχτης της Ελλάδος” [1824] and the publication “Ηπειρωτικά” (1819) by Athanasios Stageirites and then suggests that Athanasios Stageirites is the likeliest author of the “Δρομοδείχτης της Ελλάδος”.Δωδώνη: Τεύχος Πρώτο: επιστημονική επετηρίδα του Τμήματος Ιστορίας και Αρχαιολογίας της Φιλοσοφικής Σχολής του Πανεπιστημίου Ιωαννίνων; Τόμ. 43-44 (2014-2015
Dataset in support of the Southampton doctoral thesis 'The boatbuilding tradition of the Aegean during the Late Neolithic – Early Bronze Age periods. Typological classification, digital reconstruction and seakeeping assessment'
No full textDataset in support of the Southampton doctoral thesis 'The boatbuilding tradition of the Aegean during the Late Neolithic – Early Bronze Age periods. Typological classification, digital reconstruction and seakeeping assessment' Appendix D - Resistance data and Appendix C - Stability data.
This dataset is focused on two appendices:
Appendix D - Resistance data. D.1 Resistance data produced by the author via MAXSURF Resistance for this thesis.
Appendix C - Stability data
C1. Stability data – STIX and ISO criteria, produced by the author via MAXSURF Stability software for his thesis
This research was funded by Southampton Marine and Maritime Institute (SMMI), Vice-Chancellor's Scholarship, Greek Archaeological Committee UK (GACUK)
</span
Angle of Arrival Estimation Utilising Frequency Diverse Radio Antenna Arrays
No full textThe purpose of this research is to investigate a novel way of combining carrier signals that are transmitted successively over Multiple Frequencies (MF) and traditional metrics to improve AoA estimation. Every signal contains three metrics, amplitude, phase, and frequency. To achieve localisation, current systems utilise the metrics of amplitude (also known as Received Signal Strength (RSS)) and phase that resolves the AoA. However, the metric of frequency is mostly used with Orthogonal Frequency-Division Multiplexing (OFDM) to increase the number of RSS and AoA metrics, which is not optimal.
This research answers two questions. Can the use of MF improve AoA estimation? Also, how can MF and traditional metrics be combined for AoA estimation? The aim is to prove that the metric of frequency can be utilised more optimally. Therefore, measurements of RSS and AoA are performed in different environments for MF. To perform these measurements, ten frequency diverse Software Defined Radios (SDRs) are employed. A novel technique to time/frequency synchronise the SDRs is developed and presented. Moreover, a ten element Uniform Linear Array (ULA) is designed, simulated and manufactured. The outcomes of this research are two novel algorithms for the MF AoA estimation of a carrier transmitter.
Findings of the first algorithm show that the use of MF with the RSS metric performs equally with current systems that have a higher cost and complexity. The second algorithm that utilises MF with the AoA metric demonstrates a significant reduction in the AoA estimation error, compared to current systems. Specifically, for 50\% of the measured cases the AoA estimation error is reduced by 3.7 degrees, while for 95\% of the measured cases the AoA estimation error is reduced by 27 degrees. Hence, this research proves that MF with traditional metrics can reduce system complexity and greatly improve AoA estimation
Performance Evaluation of 802.15.4 UWB PHY for High Speed Data Rate under IEEE Channel Mode
No full textIn modern day society the increase of data generation and transfer has been an issue that researchers are working on. This generated and shared data might have a different purpose but one thing is certain, the reception. This communication can cover continents, countries, cities or even just a few meters. For the purpose of the later, personal area networks (PAN) have been created with a main focus to lower the energy consumption. The protocol that is created under IEEE is 802.15.4 and it has multiple applications in the context of next generation sensor networks. This thesis investigates the performance IEEE 802.15.4 UWB PHY for high data rates over IEEE multipath fading channels and introduces receivers aiming to diversity and to mitigate the intersymbol interference (ISI) that might appear. We simulate the protocols highest mandatory data rate over slow, block faded, realistic IEEE channel models such as, residential, office, outdoor and industrial. The simulation includes Reed Solomon (RS) channel coding, optimal successive erasure decoding (SED), and coherent RAKE receivers. We verify that the selective RAKE (sRAKE) perform better than the nonselective RAKE (n-sRAKE) in all environments and also the increase of fingers is mandatory in order to improve performance. In cases with low number of fingers the ISI mitigation techniques like Maximum-Likehood Sequence Estimator (MLSE) & RAKE combination or Maximum Ration Combining (MRC) ISI cancellation receivers, can provide some gain in large delay spread environments. In cases with high number of ingers the MRC received employs its full diversity since the received power is arger than before. Overall the apply of optimal errors and erasures decoding can urther improve the system performance by adding a small gain, lowering existing it Error Probability (BEP) even more.A huge percentage of data has been generated in the last two years and it will grow more, as every one of us is constantly producing and releasing data. The latest years has been an extensive research on capacity maximization, bit rate increment and power optimization. That research lead to the development of various protocols for cellular and personal area networks (PAN), where they each utilizes the frequency spectrum differently. Even if cellular networks have the ability to cover large area, development of multiple personal area networks can be developed for the purpose to offload data from the cellular network. Keeping in mind the research needs, 802.15.4 UWH PHY is a solid candidate when it comes to data transfer in a small area. This protocol offers various mandatory transmission modes that can be selected depending the channel parameters and various data rate needs. Time hopping and spreading sequence offers the existence of multiuser environment where multiple transceivers can co-exist. Overall the complexity, cost and energy consumption for transmission and reception can be kept low, matching the research needs. The main issues regarding 802.15.4 UWH PHY and high speed data rates is first, the energy dispersion of the transmitted symbol to multiple bins and second, the appearance of Inter Symbol Interference (ISI) in high delay profile environments. The solution in the former problem is the necessary implementation of a RAKE receiver. Regarding the latter, literature offers multiple ways to mitigate the ISI but the aim should be to keep the lowest complexity possible regarding the implementation. In this thesis we evaluate the performance of 802.15.4 UWB PHY for high speed data rates under IEEE channel models. Various receivers has been build for the purpose of this thesis, Maximum Ratio Combining (MRC), MRC with Inter Symbol Interference and MLSE & RAKE combination receiver. The MRC is a simple RAKE receiver with maximum diversity, MRC with ISI cancellation is based on the MRC receiver with the ability to mitigate ISI, and MLSE & RAKE combination is an optimum ISI mitigation receiver without the diversity of the MRC
Apprentissage par renforcement à champ moyen : une perspective de contrôle optimal
No full textThe goal of the PhD will be to implement a similar mean field approach to handle MARL. This idea was investigated, at least for individual agents, in several recent papers. In all of them, not only Mean field approach to MARL (Multi Agent Reinforcement Learning) does the mean field approach allow for a significant decrease of complexity, but it also provides distributed (or decentralized) solutions, which are of a very convenient use in practice. Numerical implementation using either on-or off-policy learning is discussed in the literature. The first part of the thesis will consist in revisiting the former works from a mathematical point of view. In particular, this will ask for a careful stability analysis addressing both the passage from a finite to an infinite system of agents and the use of approximated (instead of exact) policies. We may expect monotonicity to play a key role in the overall analysis; another, but more prospective, direction is to discuss the influence of a stochastic environment onto the behavior of the algorithms themselves. Another part of the thesis will be dedicated to the cooperative case the analysis of which will rely upon mean field control theory. Potential structures may allow to make the connection between individual and cooperative cases. The connection between the two may indeed play an important role for incentive design or, equivalently, for mimicking a cooperative system with individual agents. In this regard, connection with distributional reinforcement learning, may be an interesting question as wellL'apprentissage par renforcement est un paradigme clé de l'apprentissage machine, dont l'objectif est d'inciter les agents à tirer les leçons de leur propre expérience passée afin qu'ils s'améliorent au fil du temps, voir par exemple la monographie [14]. À cet égard, les systèmes impliquant un grand nombre d'agents sont importants pour les applications, mais restent difficiles à traiter du point de vue numérique, voir par exemple le récent post [12]. Le renforcement de l'apprentissage avec plusieurs agents est généralement appelé "apprentissage de renforcement multi-agents" (MARL). Comme démontré dans la publication antérieure [15], cela peut couvrir diverses situations avec des agents opérant individuellement ou collectivement. L'analyse de l'apprentissage par renforcement s'appuie fortement sur les outils mathématiques de la théorie du contrôle et de la théorie des jeux. Tout comme les MARL, les deux peuvent également être confrontés à des difficultés lorsque la dimension augmente. Cela a incité plusieurs auteurs à mettre en œuvre une approche champ moyen, issue de la physique statistique, afin de réduire la complexité globale, voir entre autres les travaux fondateurs de Lasry and Lions [9] et de Huang, Caines et Malhame [7] et les deux monographies [3, 4] sur les jeux à champ moyen et le contrôle champ moyen. L'objectif du doctorat sera de mettre en œuvre une approche similaire de la gestion des MARL. L'idée a été étudiée, au moins pour les agents individuels, dans plusieurs documents récents, voir [8, 10, 13, 16]. Dans ces derniers, non seulement l'approche champ moyen permet de réduire la complexité de façon significative, mais elle fournit également des solutions distribuées (ou décentralisées), qui sont d'une grande utilité pratique. La mise en œuvre numérique est principalement abordée dans [13, 16]. Le lien avec les notions d'apprentissage dans la théorie des jeux est cité dans [10], sur la base d'idées antérieures, voir [2]. La première partie de la thèse consistera à revisiter les travaux existants. Cela demandera en particulier une analyse soigneuse de la stabilité portant à la fois sur le passage d'un système fini d'agents à un système infini et sur l'utilisation de stratégies approximatives (au lieu de stratégies exactes). À la lumière de [2], on peut s'attendre à ce que la monotonie joue un rôle dans l'analyse globale ; une autre orientation, mais plus prospective, consiste à discuter de l'influence d'un environnement stochastique sur le comportement des algorithmes eux-mêmes. Une autre partie de la thèse sera consacrée au cas de la coopération, voir par exemple [5], dont l'analyse s'appuiera sur la théorie du contrôle en champ moyen. Comme mentionné dans [13], des structures potentielles peuvent permettre de faire le lien entre les cas individuel et coopératif ; comme démontré dans [11], ces liens jouent un rôle dans la construction de politiques incitatives
Mean field reinforcement learning : an optimal control perspective
No full textL'apprentissage par renforcement est un paradigme clé de l'apprentissage machine, dont l'objectif est d'inciter les agents à tirer les leçons de leur propre expérience passée afin qu'ils s'améliorent au fil du temps, voir par exemple la monographie [14]. À cet égard, les systèmes impliquant un grand nombre d'agents sont importants pour les applications, mais restent difficiles à traiter du point de vue numérique, voir par exemple le récent post [12]. Le renforcement de l'apprentissage avec plusieurs agents est généralement appelé "apprentissage de renforcement multi-agents" (MARL). Comme démontré dans la publication antérieure [15], cela peut couvrir diverses situations avec des agents opérant individuellement ou collectivement. L'analyse de l'apprentissage par renforcement s'appuie fortement sur les outils mathématiques de la théorie du contrôle et de la théorie des jeux. Tout comme les MARL, les deux peuvent également être confrontés à des difficultés lorsque la dimension augmente. Cela a incité plusieurs auteurs à mettre en œuvre une approche champ moyen, issue de la physique statistique, afin de réduire la complexité globale, voir entre autres les travaux fondateurs de Lasry and Lions [9] et de Huang, Caines et Malhame [7] et les deux monographies [3, 4] sur les jeux à champ moyen et le contrôle champ moyen. L'objectif du doctorat sera de mettre en œuvre une approche similaire de la gestion des MARL. L'idée a été étudiée, au moins pour les agents individuels, dans plusieurs documents récents, voir [8, 10, 13, 16]. Dans ces derniers, non seulement l'approche champ moyen permet de réduire la complexité de façon significative, mais elle fournit également des solutions distribuées (ou décentralisées), qui sont d'une grande utilité pratique. La mise en œuvre numérique est principalement abordée dans [13, 16]. Le lien avec les notions d'apprentissage dans la théorie des jeux est cité dans [10], sur la base d'idées antérieures, voir [2]. La première partie de la thèse consistera à revisiter les travaux existants. Cela demandera en particulier une analyse soigneuse de la stabilité portant à la fois sur le passage d'un système fini d'agents à un système infini et sur l'utilisation de stratégies approximatives (au lieu de stratégies exactes). À la lumière de [2], on peut s'attendre à ce que la monotonie joue un rôle dans l'analyse globale ; une autre orientation, mais plus prospective, consiste à discuter de l'influence d'un environnement stochastique sur le comportement des algorithmes eux-mêmes. Une autre partie de la thèse sera consacrée au cas de la coopération, voir par exemple [5], dont l'analyse s'appuiera sur la théorie du contrôle en champ moyen. Comme mentionné dans [13], des structures potentielles peuvent permettre de faire le lien entre les cas individuel et coopératif ; comme démontré dans [11], ces liens jouent un rôle dans la construction de politiques incitatives.The goal of the PhD will be to implement a similar mean field approach to handle MARL. This idea was investigated, at least for individual agents, in several recent papers. In all of them, not only Mean field approach to MARL (Multi Agent Reinforcement Learning) does the mean field approach allow for a significant decrease of complexity, but it also provides distributed (or decentralized) solutions, which are of a very convenient use in practice. Numerical implementation using either on-or off-policy learning is discussed in the literature. The first part of the thesis will consist in revisiting the former works from a mathematical point of view. In particular, this will ask for a careful stability analysis addressing both the passage from a finite to an infinite system of agents and the use of approximated (instead of exact) policies. We may expect monotonicity to play a key role in the overall analysis; another, but more prospective, direction is to discuss the influence of a stochastic environment onto the behavior of the algorithms themselves. Another part of the thesis will be dedicated to the cooperative case the analysis of which will rely upon mean field control theory. Potential structures may allow to make the connection between individual and cooperative cases. The connection between the two may indeed play an important role for incentive design or, equivalently, for mimicking a cooperative system with individual agents. In this regard, connection with distributional reinforcement learning, may be an interesting question as wel
Peak power reduction algorithms in asymmetric digital subscriber line modems
No full textThesis (M.Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2000.Includes bibliographical references (leaves 94-96).This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.This thesis investigates peak-to-average ratio (PAR) reduction techniques for multicarrier modulation systems, such as discrete multitone (DMT) modems and orthogonal frequency-division multiplexed (OFDM) terrestrial broadcast transmitters. Through simulation and test implementation on a state-of-the-art programmable ADSL development platform, this thesis pursues a suitable solution for minimizing PAR given the resources of a programmable platform. This solution is integrated as a prototype implementation into a fully-functional ADSL modem and optimized for maximum PAR reduction performance within modem complexity constraints.by Athanasios Dimitri Dousis.M.Eng
Homophobic Statements, a Bishop, and the Limits of Freedom of Expression. An In-Depth Commentary on ECtHR 31.08.2023, Amvrosios-Athanasios Lenis v. Greece, no. 47833/20
No full textDichiarazioni omofobe, un vescovo e i limiti della libertà di espressione. Un commento approfondito su CEDU 31.08.2023, Amvrosios-Athanasios Lenis v. Greece, no. 47833/20.
ABSTRACT: The decision of the ECtHR of 31.08.2023, Amvrosios-Athanasios Lenis v. Greece (no. 47833/20), is a further step toward an increasingly dense jurisprudence on “hate speech” and the limits of freedom of expression. The public proclamation of religious doctrines that are in conflict with the values of the contracting States enshrined in the ECHR is protected to a certain extent by the fundamental right of freedom of religion and belief. However, the qualification of a statement as religious does not justify “hate speech.” The AUTHOR shows the tension between freedom of religion, freedom of expression, and protection against discrimination, and analyzes the decision against the backdrop of Article 17 of the ECHR (prohibition of abuse of rights).
SOMMARIO: 1. Preliminary Remarks - 2. The Concept of Hate Speech - 3. The Facts of the Case - 4. The Procedure and Reasoning of the Court - 4.1 The ECtHR’s Preliminary Considerations on Fundamental Rights - 4.2 Legal assessment - 4.3 Some Remarks on (the Non-Invoked) Article 9 of the ECHR - 5. Concluding Remarks
An introduction to Mean Field Games using probabilistic methods
No full textΕθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) “Μαθηματική Προτυποποίηση σε Σύγχρονες Τεχνολογίες και στα Χρηματοοικονομικά
- …
