KMi at Web of Data summer schoolChris Valentine, Wednesday 24 August 2011 | Annotate
KMi professor Stefan Rueger taught international students last week at the summer school "Web of Data" held in St Petersburg, Russia. The summer school aimed to bring cutting-edge topics, problems and methods in Information Retrieval and Database Technology to graduate and PhD students as well as experienced academic and industrial researchers. The main sponsors of the summer school were Yandex, a Russian search engine company, and Google.
Stefan Rueger gave a series of lectures on multimedia information retrieval and discussed new and emerging research topics with the participants. Multimedia databases are amongst the most difficult databases to query and index meaningfully. This course looked at cases where the query itself is a multimedia excerpt, for example, when you stroll around in an unknown place and want to send a photo of an interesting landmark to a search engine as a query hoping you are told more about it. This course discussed underlying techniques and common approaches to facilitate multimedia queries and indexing: fingerprinting of multimedia to recognise known objects; metadata and piggy-back text access where automated processes create text surrogates for multimedia; automated image annotation; content-based image queries and indexing. Thelatter was studied in great depth analysing features and distances, and how to effectively combine them for efficient database access.
All in all eight topics were covered during the week of the summer school in two parallel sessions, while the evenings between 6pm and 7.30pm were taken by an adjoining Young Scientists Conference in Information Retrieval. Four Social events from 8pm to midnight ensured there was every opportunity to network beyond learning. The school was co-organized by Saint Petersburg State University, Russian Information Retrieval Evaluation Seminar (ROMIP), and EDBT Association.