Speech processing, the field of analysing input speech signals and methods of processing them has emerged in the recent days. Additionally, the development of a speech processing system involves several components in the design phase with probabilistic approximations for enhanced audio sampling and de-noising. In this work, we focus into use of Gaussian random variables while modelling and filtering noise that gets added after being passed through an additive noise channel in a communication system, and the applications of Hidden Markov models. Moreover, we apply deep learning methods for emotion classification via a robust and accurate ensemble learning scheme that is applied to a joint deep network which incorporates audiovisual inputs and generates the emotion prediction effectively reaching satisfactory accuracy.

At the dawn of the Internet era graph analytics play an important role in high- and low-level network policymaking across a wide array of fields so diverse as transportation network design, supply chain engineering and logistics, social media analysis, and computer communication networks, to name just a few. This can be attributed not only to the size of the original graph but also to the nature of the problem parameters. For instance, algorithmic solutions depend heavily on the approximation criterion selection. Moreover, iterative or heuristic solutions are often sought as it is a high dimensional problem given the high number of vertices and edges involved as well as their complex interaction. Replacing under constraints a directed graph with an undirected one having the same vertex set is often sought in applications such as data visualization, community structure discovery, and connection-based vertex centrality metrics. Polar decomposition is a key matrix factorization which represents a matrix as a product of a symmetric positive (semi)definite factor and an orthogonal one. The former can be an undirected approximation of the original adjacency matrix. The proposed graph approximation has been tested with three Twitter graphs with encouraging results with respect to density, Fiedler number, and certain vertex centrality metrics based on matrix power series. The dataset was hosted in an online MongoDB instance.

Graph neural networks (GNNs) is an emerging class of iterative connectionist models taking full advantage of the interaction patterns in an underlying domain. Depending on their configuration GNNs aggregate local state information to obtain robust estimates of global properties. Since graphs inherently represent high dimensional data, GNNs can effectively perform dimensionality reduction for certain aggregator selections. One such task is assigning sentiment polarity labels to the vertices of a large social network based on local ground truth state vectors containing structural, functional, and affective attributes. Emotions have been long identified as key factors in the overall social network resiliency and determining such labels robustly would be a major indicator of it. As a concrete example, the proposed methodology has been applied to two benchmark graphs obtained from political Twitter with topic sampling regarding the Greek 1821 Independence Revolution and the US 2020 Presidential Elections. Based on the results recommendations for researchers and practitioners are offered.

Suppose we have a virus or one competing idea/product that propagates over a multiple profile (e.g., social) network. Can we predict what proportion of the network will actually get "infected" (e.g., spread the idea or buy the competing product), when the nodes of the network appear to have different sensitivity based on their profile? For example, if there are two profiles A and B in a network and the nodes of profile A and profile B are susceptible to a highly spreading virus with probabilities βA and βB respectively, what percentage of both profiles will actually get infected from the virus at the end? To reverse the question, what are the necessary conditions so that a predefined percentage of the network is infected? We assume that nodes of different profiles can infect one another and we prove that under realistic conditions, apart from the weak profile (great sensitivity), the stronger profile (low sensitivity) will get infected as well. First, we focus on cliques with the goal to provide exact theoretical results as well as to get some intuition as to how a virus affects such a multiple profile network. Then, we move to the theoretical analysis of arbitrary networks. We provide bounds on certain properties of the network based on the probabilities of infection of each node in it when it reaches the steady state. Finally, we provide extensive experimental results that verify our theoretical results and at the same time provide more insight on the problem.

Positioning capabilities offered in modern mobile devices enable usage of location-based services. Privacy and security is of great importance for related applications. We present a framework that allows conducting research on anonymity techniques in a real-life environment using smartphones. The proposed solution also includes logging mechanisms that facilitate positioning research dataset development in open format. To present the capabilities of the solution, we deliver the concept of K-anonymity to protect mobile users that issue queries to location-based services. Experimental evaluation of the solution includes development of real-life logging dataset using smartphones by volunteers. Different flavours of anonymity algorithms are easy to be included and tested. The solution has received encouraging feedback and successfully assists the researchers of location based services to experiment, validate and develop their techniques in real life environment.

NoSQL databases focus on analytical processing of large scale datasets, offering increased scalability over commodity hardware. One of their strongest features is elasticity, which allows for fairly portioned premiums and high-quality performance. Yet, the process of adaptive expansion and contraction of resources usually involves a lot of manual effort, often requiring the definition of the conditions for scaling up or down to be provided by the users. To date, there exists no open-source system for automatic resizing of NoSQL clusters. In this demonstration, we present TIRAMOLA, a modular, cloud-enabled framework for monitoring and adaptively resizing NoSQL clusters. Our system incorporates a decision-making module which allows for optimal cluster resize actions in order to maximize any quantifiable reward function provided together with life-long adaptation to workload or infrastructural changes. The audience will be able to initiate HBase clusters of various sizes and apply varying workloads through multiple YCSB clients. The attendees will be able to watch, in real-time, the system perform automatic VM additions and removals as well as how cluster performance metrics change relative to the optimization parameters of their choice.

This work studies the problem of 2-dimensional searching for the 3-sided range query of the form [a, b] x (-∞, c] in both main and external memory, by considering a variety of input distributions. A dynamic linear main memory solution is proposed, which answers 3-sided queries in O(log n + t) worst case time and scales with O (log log n) expected with high probability update time, under continuous μ-random distributions of the x and y coordinates, where n is the current number of stored points and t is the size of the query output. Our expected update bound constitutes a considerable improvement over the O(log n) update time bound achieved by the classic Priority Search Tree of McCreight [23], as well as over the Fusion Priority Search Tree of Willard [30], which requires O(log n/log log n) time for all operations. Moreover, we externalize this solution, gaining O(logB n + t/B) worst case and O(logBlogn) amortized expected with high probability I/Os for query and update operations respectively, where B is the disk block size. Then, combining the Modified Priority Search Tree [27] with the Priority Search Tree [23], we achieve a query time of O(log log n + t) expected with high probability and an update time of O(log log n) expected with high probability, under the assumption that the x-coordinates are continuously drawn from a smooth distribution and the y-coordinates are continuously drawn from a more restricted class of distributions. The total space is linear. Finally, we externalize this solution, obtaining a dynamic data structure that answers 3-sided queries in O(logB log n + t/B) I/Os expected with high probability, and it can be updated in O(logB log n) I/Os amortized expected with high probability and consumes O(n/B) space, under the same assumptions.

In this paper we introduce a novel distributed simulation environment with GUI for P2P simulations (D-P2P-Sim). The key aim is to provide the appropriate integrated set of tools in a single software solution to evaluate the performance of various protocols. The basic architecture of the distributed P2P simulator is based on a multi-threading, asynchronous, message passing and distributed environment with graphical user interface to facilitate ease of use by both researchers and programmers.