Young Researchers!

Substantial volumes of geo-tagged text content are available, including web pages, business directory entries, and microblog posts. Such geo-textual web content is important to many human activities. Thus, studies suggest that each week, several billion keyword-based queries are issued that have some form of local intent. Such queries aim to find relevant and nearby geo-textual content. This state of affairs renders it important to be able to efficiently compute queries that retrieve useful content from large collections of geo-textual web content. A standard keyword-based query on geo-textual content takes a user location and user-supplied keywords as arguments, and it returns content that is geographically near and textually relevant to these arguments. Due perhaps to the rich semantics of geographical space and its importance to our daily lives, many different kinds of useful geo-textual web query functionality may be envisioned. For example, some queries aim to find a small set of near-by points of interest where each point of interest satisfies a user’s needs; other queries aim to find a set of points of interest that collectively satisfy a user’s needs; and yet other queries aim to find regions that are densely populated by relevant points of interest, thus enabling a user to conveniently explore relevant points of interest. Based on recent and ongoing work by the speaker and his colleagues, the talk presents key functionality, concepts, and techniques relating to the querying of geo-textual web content; it covers functionality that addresses different kinds of user intent; and it offers directions for the future development of keyword-based geo-textual web querying.

Many searches on the Web suffer from a problem often referred to as underspecified queries where the user provides the least possible constraints and the search system has to fill the rest. Progress is made over the past two decades addressing different aspects of the problem, but important challenges still remain open. In particular, some fundamental questions which are not well-studied are when and for what classes of searches an underspecified query can be resolved and how the choice of a resolution affects the search and the kinds of resources that can be retrieved.
In this talk, I will touch on some of these issues, reviewing some of the past research in this area with a focus on some of our own work; I will highlight some underlying research issues that are open and a few areas of possible future research directions.

In this talk, we discuss some interesting research issues related to online social networks, in particular, influenceability evaluation of nodes and random-walk domination over social networks. The influenceability evaluation is an important issue in influence analysis. The important applications of the influence analysis in social network include influence maximization for viral marketing,finding the most influential nodes, online advertising, etc. For many of these applications, it is crucial to evaluate the influenceability of a node. In this talk, we discuss evaluating influenceability of nodes in social network based on the widely used influence spread model, namely, the independent cascade model. From a different viewpoint, we discuss random-walk domination problems in social networks motivated by a number of applications in practice (e.g., item-placement problem in online social networks, Ads-placement problem in advertisement networks, and resource-placement problem in P2P networks).

Microblogs, e.g., tweets, blogs or news comments, are witnessing flourishing era with the widespread of mobile devices and users. Such microblogs form a stream of user-generated data with rich attributes that arrive with very high arrival rates of thousands of microblogs every second. Due to its richness, microblogs have been exploited for various applications that include news extraction, event detection, understanding public opinions, targeted advertising, among others. However, most of the efforts of dealing with microblogs were sporadic in the sense that various algorithms have been proposed here and there to solve particular problems. In this talk, i will advocate for the need of having a full-fledged system that fully supports microblogs data management issues. The talk will go through the details of the envisioned systems and the various components needed to achieve the most common microblogs functionality. The talk will be concluded by showing a live demo of the TAGHREED system that presents our first attempt for realizing a Microblog Data Management System.