I am one of the organizers of the „Spoken Web Search“ Task at MediaEval 2011, which aims to build a language-independent speech retrieval system with very little resources.

From the task description at http://www.multimediaeval.org/mediaeval2011/SWS2011/:

The task involves searching FOR audio content WITHIN audio content USING an audio content query. This task is particularly interesting for speech researchers in the area of spoken term detection. The task addresses the challenge of multiple, resource-limited languages. The task would require researchers to build a language-independent audio search system so that given a query should be able to find the appropriate audio file and the location of query term within the audio file.