Sergey Chernov

Exploiting Flickr Tags and Groups for Finding Landmark Photos

Many people take pictures of different city landmarks and post them to photo-sharing systems like Flickr. They also add tags and place photos in Flickr groups, created around particular themes. Using tags, other people can search for representative landmark images of places of interest. Searching for landmarks using tags results into many non-landmark photos and provides poor landmark summary for a city. In this paper we propose a new method to identify landmark photos using tags and social Flickr groups. In contrast to similar modern systems, our approach is also applicable when GPS-coordinates for photos are not available. Presented user study shows that the proposed method outperforms state-of-the-art systems for landmark finding.

Building a desktop search test-bed

In the last years several top-quality papers utilized temporary Desktop data and/or browsing activity logs for experimental evaluation. Building a common testbed for the Personal Information Management community is thus becoming an indispensable task. In this paper we present a possible dataset design and discuss the means to create it.

Task detection for activity-based desktop search

The desktop search tools provide powerful query capabilities and result presentation techniques. However, they do not take the user context into account. We propose to exploit collected information about user activities with desktop files and applications for activity-based desktop search. When I prepare for a project review and type in a search box the name of a colleague, I expect to find her last deliverable draft, but not her email with a paper review or our joint conference presentation. Ideally, the desktop search system should be able to infer my current task from the logs of my previous activities and present task-specific search results.

Enhancing expert search through query modeling

An expert finding is a very common task among enterprise search activities, while its usual retrieval performance is far from the quality of the Web search. Query modeling helps to improve traditional document retrieval, so we propose to apply it in a new setting. We adopt a general framework of language modeling for expert finding. We show how expert language models can be used for advanced query modeling. A preliminary experimental evaluation on TREC Enterprise Track 2006 collection shows that our method improves the retrieval precision on the expert finding task.

A plugin architecture enabling federated search for digital libraries

Today, users expect a variety of digital libraries to be searchable from a single Web page. The German Vascoda project provides this service for dozens of information sources. Its ultimate goal is to provide search quality close to the ranking of a central database containing documents from all participating libraries. Currently, however, the Vascoda portal is based on a non-cooperative metasearch approach, where results from sources are merged randomly and ranking quality is sub-optimal. In this paper, we describe a Lucene-based plugin which replaces this method by a truly federated search across different search engines, where the exchange of document statistics improves document ranking. Preliminary evaluation results show ranking results equal to a centralized setup.

GuideMe! The World of sights in your pocket

Web 2.0 applications are a rich source of multimedia resources, that describe sights, events, whether conditions, traffic situations and other relevant objects along the users route. Compared to static sight descriptions, Web 2.0 resources can provide up-to-date visual information, which has been found important or interesting by the other users. Some algorithms have been suggested recently for the landmark finding problem from photos. Still, if users want related videos or background information about a particular place of interest it is necessary to contact different social platforms or general search engines. In this paper we present GuideMe! - a mobile application that automatically identifies landmark tags from Flickr groups and gathers relevant sightseeing resources from various Web 2.0 social platforms.