Ilias Kalamaras

Multi-Objective Optimization for Multimodal Visualization

Using data visualization techniques can be of significant assistance in exploring multimedia databases. Data visualization is typically addressed as a unimodal learning task, where data are described with only one feature set, or modality. However, using multiple data modalities has been proved to increase the performance of learning methods. In this paper a novel approach for exploiting the multiple available modalities for visualization is proposed, motivated by the field of multi-objective optimization. Initially, each modality is considered separately. A graph of the dissimilarities among the data and the corresponding minimum spanning tree are formed. The suitability of a particular data placement is quantified using multiple cost functions, one for each modality. The utilized cost functions are defined in terms of graph aesthetic measures, computed for the unimodal minimum spanning trees. The cost functions are then used as the multiple objectives of a multi-objective optimization problem. Solving the problem results in a set of Pareto optimal placements, which represent different trade-offs among the various objectives. Experimental evaluation shows that the proposed method outperforms current multimodal visualization methods both in discovering more visualizations and in producing ones which are more aesthetically pleasing and easily perceivable.

An enhanced Graph Analytics Platform (GAP) providing insight in Big Network Data

Abstract Being a widely adapted and acknowledged practice for the representation of inter- and intra-dependent information streams, network graphs are nowadays growing vast in both size and complexity, due to the rapid expansion of sources, types, and amounts of produced data. In this context, the efficient processing of the big amounts of information, also known as Big Data forms a major challenge for both the research community and a wide variety of industrial sectors, involving security, health and financial applications. Serving these emerging needs, the current paper presents a Graph Analytics based Platform (GAP) that implements a top-down approach for the facilitation of Data Mining processes through the incorporation of state-of-the-art techniques, like behavioural clustering, interactive visualizations, multi-objective optimization, etc. The applicability of this platform is validated on 2 istinct real-world use cases, which can be considered as characteristic examples of modern Big Data problems, due to the vast amount of information they deal with. In particular, (i) the root cause analysis of a Denial of Service attack in the network of a mobile operator and (ii) the early detection of an emerging event or a hot topic in social media communities. In order to address the large volume of the data, the proposed application starts with an aggregated overview of the whole network and allows the operator to gradually focus on smaller sets of data, using different levels of abstraction. The proposed platform offers differentiation between different user behaviors that enable the analyst to obtain insight on the network’s operation and to extract the meaningful information in an effortless manner. Dynamic hypothesis formulation techniques exploited by graph traversing and pattern mining, enable the analyst to set concrete network-related hypotheses, and validate or reject them accordingly.

MoVA: A Visual Analytics Tool Providing Insight in the Big Mobile Network Data

Mobile networks have numerous exploitable vulnerabilities that enable malicious individuals to launch Denial of Service (DoS) attacks and affect network security and performance. The efficient detection and attribution of these anomalies are of major importance to the mobile network operators, especially since there is a vast amount of information collected, which renders the problem as a Big Data problem. Previous approaches focus on either anomaly detection methods, or visualization methods separately. In addition, they utilize solely either the signaling or the Call Detail Record (CDR) activity in the network. This paper presents MoVA (Mobile network Visual Analytics), a visual analytics tool for the detection and attribution of anomalies in mobile cellular networks which combines anomaly detection and visualization, and is applied on both signaling and CDR activity in the network. In order to address the large volume of the data, the proposed application starts with an aggregated overview of the whole network and allows the operator to gradually focus on smaller sets of data, using different levels of abstraction. The proposed visualization methods are able to differentiate between different user behaviors, and enable the analyst to have an insight in the mobile network operation and easily spot the anomalous mobile devices. Hypothesis formulation and validation methods are also provided, in order to enable the analyst to formulate network security-related hypotheses, and validate or reject them based on the results of the analysis.

A multi-objective clustering approach for the detection of abnormal behaviors in mobile networks

The visualization of mobile network data can be of significant value to the network security administrator in order to detect anomalies in the normal traffic, caused by malicious attacks. Although several visualization types of the network structure and traffic already exist, the literature around visualizing behavioral aspects of users or network components, in order to distinguish the normal from the abnormal ones, is limited. In this paper, a behavior-based approach for visualizing the users of the network, with respect to specific aspects of their behavior, is proposed. The approach introduces the extraction of behavior-related descriptors from the raw network traffic data, which can be used to visualize behavioral similarities, so that users with similar behavior are depicted as points close to each other. Multiple descriptors are extracted from each user and are used as the multiple modalities in a state-of-the-art multi-objective visualization method. The outcome of the multi-objective method is a visualization of the behavioral similarities of users, according to the selection of a trade-off among the multiple descriptors. This allows the analyst to visually detect anomalies and analyze their evolution in time. Experimental evaluation of the proposed approach with several datasets in various application scenarios verify its efficiency.

Towards visualizing primary Sjögren's Syndrome data from heterogeneous cohorts

Visual analytics establish a comprehensive approach to handling the exponential growth of healthcare data and promise to offer innovative approaches to the understanding of health parameters and their important interrelations. Primary Sjögrens Syndrome (pSS) is an autoimmune disease with unknown causes and various symptoms, for which visual analytics can prove useful in understanding its characteristics, utilizing the abundance of currently available data. A large number of medical organizations currently possess databases of patients with pSS, recording demographic, geographical, clinical, genetic and activity data. However, these databases are usually diverse in their schemas, focusing on different characteristics and having different naming conventions for their concepts. Visual analytics for such data require that they are represented in a common schema. This paper presents the Visual analytics methods utilized within the HarmonicSS EU project, which aim at providing visualization and interaction techniques to the operator, based on a semantic-based harmonization of data from multiple cohorts. Visualization of large data from multiple sources is important in order to understand the causes of the disease and facilitate diagnosis.

Towards Big Data Analytics in Large-Scale Federations of Semantically Heterogeneous IoT Platforms

The technological advances in the Internet-of-Things (IoT) have led to the generation of large amounts of data and the production of a large number of IoT platforms for their management. The abundance of raw data necessitates the use of data analytics in order to extract useful patterns for decision making. Current architectures for big data analytics in the IoT domain address the large volume and velocity of the produced data. However, they do not address the semantic heterogeneity in the data models used by diverse IoT platforms, which emerges when large-scale deployments, spanning across multiple deployment sites, are considered. This paper proposes an architecture for big data analytics in the context of large-scale IoT systems consisting of multiple IoT platforms. A Semantic Interoperability Layer (SIL) handles the interoperability among the data models of the individual platforms, using semantic mappings between them and a unified ontology. Data queries to the SIL and result collection is handled by a cloud-based data management layer, namely the Data Lake, along with storage of metadata needed by data analytics methods. Based on this infrastructure, web-based data analytics and visual analytics methods are used to analyze the collected data, while being agnostic of platform-specific details. The proposed architecture is developed in the context of healthcare provision for older people, although it can be applied to any IoT domain.

A Multi-objective Data Mining Approach for Road Traffic Prediction

Road traffic prediction for the efficient traffic control has lately been in the focus of the research community, as it can solve significant urban issues, such as city evacuation plans, increased concentration of CO2 emissions and delays caused by extended traffic jams. The current paper proposes a novel approach for multi-variate data mining from past traffic data (i.e. average speed values per road), so as to dynamically detect all significant correlations between the road network components (i.e. the segments of the roads) by mapping the latter onto a low dimensional embedding. Multiple traffic-related features (e.g. speed correlation, spatial proximity, phase difference, etc.) are utilized in a multi-objective optimization framework, producing all Pareto-optimal embeddings, each one corresponding to a different trade-off between the objectives. The operator is provided with the option to interactively select among these Pareto-optimal solutions, so as to explore the most descriptive sets of road influences. The proposed method has been evaluated on real traffic data, while the evaluation of the forecasting performance of the multi-objective approach exhibited accuracy improvement with respect single-objective approaches.

Accessibility-based reranking in multimedia search engines

Traditional multimedia search engines retrieve results based mostly on the query submitted by the user, or using a log of previous searches to provide personalized results, while not considering the accessibility of the results for users with vision or other types of impairments. In this paper, a novel approach is presented which incorporates the accessibility of images for users with various vision impairments, such as color blindness, cataract and glaucoma, in order to rerank the results of an image search engine. The accessibility of individual images is measured through the use of vision simulation filters. Multi-objective optimization techniques utilizing the image accessibility scores are used to handle users with multiple vision impairments, while the impairment profile of a specific user is used to select one from the Pareto-optimal solutions. The proposed approach has been tested with two image datasets, using both simulated and real impaired users, and the results verify its applicability. Although the proposed method has been used for vision accessibility-based reranking, it can also be extended for other types of personalization context.