Deep bottleneck network based i-vector representation for language identification.
Yan SongXinhai HongBing JiangRuilian CuiIan McLoughlinLi-Rong DaiPublished in: INTERSPEECH (2015)
Keyphrases
- language identification
- vector representation
- similarity measure
- fixed length
- document representation
- similarity computation
- speaker identification
- semantic description
- distance measure
- document images
- web information retrieval
- information retrieval
- vector space
- speech recognition
- document collections
- web search
- feature vectors
- natural language
- multiscale
- web pages