arXiv subjects

Evaggelia Pitoura

Publications and source records attributed to Evaggelia Pitoura.

31 records · Page 2Linked to original sources

arXiv2021

Fairness-Aware PageRank

Algorithmic fairness has attracted significant attention in the past years. Surprisingly, there is little work on fairness in networks. In this work, we consider fairness for link analysis algorithms and in particular for the celebrated PageRank algorithm. We provide definitions for fairness, and propose two approaches for achieving fairness. The first modifies the jump vector of the Pagerank algorithm to enfonce fairness, and the second imposes a fair behavior per node. We also consider the problem of achieving fairness while minimizing the utility loss with respect to the original algorithm. We present experiments with real and synthetic graphs that examine the fairness of Pagerank and demonstrate qualitatively and quantitatively the properties of our algorithms.

cs.SI↗

arXiv2020

Flow Computation in Temporal Interaction Networks

Temporal interaction networks capture the history of activities between entities along a timeline. At each interaction, some quantity of data (money, information, kbytes, etc.) flows from one vertex of the network to another. Flow-based analysis can reveal important information. For instance, financial intelligent units (FIUs) are interested in finding subgraphs in transactions networks with significant flow of money transfers. In this paper, we introduce the flow computation problem in an interaction network or a subgraph thereof. We propose and study two models of flow computation, one based on a greedy flow transfer assumption and one that finds the maximum possible flow. We show that the greedy flow computation problem can be easily solved by a single scan of the interactions in time order. For the harder maximum flow problem, we propose graph precomputation and simplification approaches that can greatly reduce its complexity in practice. As an application of flow computation, we formulate and solve the problem of flow pattern search, where, given a graph pattern, the objective is to find its instances and their flows in a large interaction network. We evaluate our algorithms using real datasets. The results show that the techniques proposed in this paper can greatly reduce the cost of flow computation and pattern enumeration.

cs.DB↗

arXiv2020

A Hybrid Approach to Temporal Pattern Matching

The primary objective of graph pattern matching is to find all appearances of an input graph pattern query in a large data graph. Such appearances are called matches. In this paper, we are interested in finding matches of interaction patterns in temporal graphs. To this end, we propose a hybrid approach that achieves effective filtering of potential matches based both on structure and time. Our approach exploits a graph representation where edges are ordered by time. We present experiments with real datasets that illustrate the efficiency of our approach.

cs.DS↗

arXiv2020

Forming Compatible Teams in Signed Networks

The problem of team formation in a social network asks for a set of individuals who not only have the required skills to perform a task but who can also communicate effectively with each other. Existing work assumes that all links in a social network are positive, that is, they indicate friendship or collaboration between individuals. However, it is often the case that the network is signed, that is, it contains both positive and negative links, corresponding to friend and foe relationships. Building on the concept of structural balance, we provide definitions of compatibility between pairs of users in a signed network, and algorithms for computing it. We then define the team formation problem in signed networks, where we ask for a compatible team of individuals that can perform a task with small communication cost. We show that the problem is NP-hard even when there are no communication cost constraints, and we provide heuristic algorithms for solving it. We present experimental results with real data to investigate the properties of the different compatibility definitions, and the effectiveness of our algorithms.

cs.DS↗

arXiv2018

Bias Disparity in Recommendation Systems

Recommender systems have been applied successfully in a number of different domains, such as, entertainment, commerce, and employment. Their success lies in their ability to exploit the collective behavior of users in order to deliver highly targeted, personalized recommendations. Given that recommenders learn from user preferences, they incorporate different biases that users exhibit in the input data. More importantly, there are cases where recommenders may amplify such biases, leading to the phenomenon of bias disparity. In this short paper, we present a preliminary experimental study on synthetic data, where we investigate different conditions under which a recommender exhibits bias disparity, and the long-term effect of recommendations on data bias. We also consider a simple re-ranking algorithm for reducing bias disparity, and present some observations for data disparity on real data.

cs.IR↗

arXiv2018

Flow Motifs in Interaction Networks

Many real-world phenomena are best represented as interaction networks with dynamic structures (e.g., transaction networks, social networks, traffic networks). Interaction networks capture flow of data which is transferred between their vertices along a timeline. Analyzing such networks is crucial toward comprehend- ing processes in them. A typical analysis task is the finding of motifs, which are small subgraph patterns that repeat themselves in the network. In this paper, we introduce network flow motifs, a novel type of motifs that model significant flow transfer among a set of vertices within a constrained time window. We design an algorithm for identifying flow motif instances in a large graph. Our algorithm can be easily adapted to find the top-k instances of maximal flow. In addition, we design a dynamic programming module that finds the instance with the maximum flow. We evaluate the performance of the algorithm on three real datasets and identify flow motifs which are significant for these graphs. Our results show that our algorithm is scalable and that the real networks indeed include interesting motifs, which appear much more frequently than in randomly generated networks having similar characteristics.

cs.SI↗

arXiv2017

On Measuring Bias in Online Information

Bias in online information has recently become a pressing issue, with search engines, social networks and recommendation services being accused of exhibiting some form of bias. In this vision paper, we make the case for a systematic approach towards measuring bias. To this end, we discuss formal measures for quantifying the various types of bias, we outline the system components necessary for realizing them, and we highlight the related research challenges and open problems.

cs.DB↗

arXiv2017

Best Friends Forever (BFF): Finding Lasting Dense Subgraphs

Graphs form a natural model for relationships and interactions between entities, for example, between people in social and cooperation networks, servers in computer networks, or tags and words in documents and tweets. But, which of these relationships or interactions are the most lasting ones? In this paper, we study the following problem: given a set of graph snapshots, which may correspond to the state of an evolving graph at different time instances, identify the set of nodes that are the most densely connected in all snapshots. We call this problem the Best Friends For Ever (BFF) problem. We provide definitions for density over multiple graph snapshots, that capture different semantics of connectedness over time, and we study the corresponding variants of the BFF problem. We then look at the On-Off BFF (O^2BFF) problem that relaxes the requirement of nodes being connected in all snapshots, and asks for the densest set of nodes in at least $k$ of a given set of graph snapshots. We show that this problem is NP-complete for all definitions of density, and we propose a set of efficient algorithms. Finally, we present experiments with synthetic and real datasets that show both the efficiency of our algorithms and the usefulness of the BFF and the O^2BFF problems.

cs.SI↗

arXiv2016

Boosting Nodes for Improving the Spread of Influence

Information diffusion in networks has received a lot of recent attention. Most previous work addresses the influence maximization problem of selecting an appropriate set of seed nodes to initiate the diffusion process so that the largest number of nodes is reached. Since the seed selection problem is NP hard, most solutions are sub-optimal. Furthermore, there may be settings in which the seed nodes are predetermined. Thus, a natural question that arise is: given a set of seed nodes, can we select a small set of nodes such that if we improve their reaction to the diffusion process, the largest increase in diffusion spread is achieved? We call this problem, the boost set selection problem. In this paper, we formalize this problem, study its complexity and propose appropriate algorithms. We also evaluate the effect of boosting in a number of real networks and report the increase of influence spread achieved for different seed sets, time limits in the diffusion process and other diffusion parameters.

cs.SI↗

arXiv2014

Rapid AkNN Query Processing for Fast Classification of Multidimensional Data in the Cloud

A $k$-nearest neighbor ($k$NN) query determines the $k$ nearest points, using distance metrics, from a specific location. An all $k$-nearest neighbor (A$k$NN) query constitutes a variation of a $k$NN query and retrieves the $k$ nearest points for each point inside a database. Their main usage resonates in spatial databases and they consist the backbone of many location-based applications and not only (i.e. $k$NN joins in databases, classification in data mining). So, it is very crucial to develop methods that answer them efficiently. In this work, we propose a novel method for classifying multidimensional data using an A$k$NN algorithm in the MapReduce framework. Our approach exploits space decomposition techniques for processing the classification procedure in a parallel and distributed manner. To our knowledge, we are the first to study the classification of multidimensional objects under this perspective. Through an extensive experimental evaluation we prove that our solution is efficient and scalable in processing the given queries. We investigate many different perspectives that can affect the total computational cost, such as different dataset distributions, number of dimensions, growth of $k$ value and granularity of space decomposition and prove that our system is efficient, robust and scalable.

cs.DB↗

arXiv2013

DisC Diversity: Result Diversification based on Dissimilarity and Coverage

Recently, result diversification has attracted a lot of attention as a means to improve the quality of results retrieved by user queries. In this paper, we propose a new, intuitive definition of diversity called DisC diversity. A DisC diverse subset of a query result contains objects such that each object in the result is represented by a similar object in the diverse subset and the objects in the diverse subset are dissimilar to each other. We show that locating a minimum DisC diverse subset is an NP-hard problem and provide heuristics for its approximation. We also propose adapting DisC diverse subsets to a different degree of diversification. We call this operation zooming. We present efficient implementations of our algorithms based on the M-tree, a spatial index structure, and experimentally evaluate their performance.

cs.DB↗

arXiv2013

Finding the Right Set of Users: Generalized Constraints for Group Recommendations

Recently, group recommendations have attracted considerable attention. Rather than recommending items to individual users, group recommenders recommend items to groups of users. In this position paper, we introduce the problem of forming an appropriate group of users to recommend an item when constraints apply to the members of the group. We present a formal model of the problem and an algorithm for its solution. Finally, we identify several directions for future work.

cs.IR↗

arXiv2013

On Graph Deltas for Historical Queries

In this paper, we address the problem of evaluating historical queries on graphs. To this end, we investigate the use of graph deltas, i.e., a log of time-annotated graph operations. Our storage model maintains the current graph snapshot and the delta. We reconstruct past snapshots by applying appropriate parts of the graph delta on the current snapshot. Query evaluation proceeds on the reconstructed snapshots but we also propose algorithms based mostly on deltas for efficiency. We introduce various techniques for improving performance, including materializing intermediate snapshots, partial reconstruction and indexing deltas.

cs.DB↗