Translating text to queries for audiovisual content recommendation

Title:Translating text to queries for audiovisual content recommendation
Institute:University of Twente (HMI)
Place:Enschede The Netherlands
Type:Capita selecta and Research Topics
End date:not present
HMI ContactRoeland OrdelmanRobin Aly



There are many interesting use cases for video recommendation based on rich text sources such as news articles, blogs or recipes that are available to the Internet. For example, one important use case is recommending videos to users based on the news feed they are reading. A key challenge is to translate the text source into a query that can be used for searching a video collection. The query should represent the contents of the text in a proper way but also take into account properties of the video collection, e.g. the availability of metadata, and the abilities of the search engine, e.g. the ability to search for the appearance of persons.


In the above context, we offer a short-research assignment that consists of a study on related work and the evaluation of a baseline system for query translation.

NB: we also offer a final project on query translation for video recommendation, focusing on the development and evaluation of advanced methods.