Click here to visit the SIG website!
Aim. The primary goal of SLIM is to bring together all protagonists working in speech and language processing to analyze, index and access multimedia data. Indeed, the SLIM SIG seeks to build a strong and well structured community on the topic to: strengthen the existing research efforts; foster and promote new research directions; develop new and fruitful collaborations; gain visibility in the speech and multimedia communities.
In general, the SLIM SIG is of interest to people doing speech and language related research on any type of multimedia material. In particular, the main areas that the SIG covers include (but are not limited to):
Motivation. All aspects of spoken language - pragmatic, semantic, acoustic, phonetic, speaker, etc. - clearly play a major role in fine-grain semantic analysis and access of multimedia data such as lectures, meetings, interviews, debates, conversational broadcast, podcasts, social videos on the Web, etc. However, such data, along with the associated use scenarios, raise specific challenges to speech technologies. For example, high transcription error rates can be observed due to the sometime poor quality of the data; semantics can be shared between several modalities; etc. Bringing together researchers from across our community allows us to direct focused effort at addressing these challenges.
Board. Current SIG officers are:
The SLAM Special Interest Group supports the series of workshops on Speech, Language and Audio in Multimedia (SLAM)
SLAM also supports the MediaEval benchmark initiative