The use of semantic and acoustic features for open-domain TED talk summarization.
Fajri KotoSakriani SaktiGraham NeubigTomoki TodaMirna AdrianiSatoshi NakamuraPublished in: APSIPA (2014)
Keyphrases
- open domain
- acoustic features
- information extraction
- question answering
- speaker verification
- question answering systems
- speech signal
- music information retrieval
- automatic speech recognition
- visual features
- natural language
- semantic information
- semantic similarity
- semantic features
- high level
- feature selection
- noisy environments
- semantically related
- co occurrence
- image retrieval