Disfluency Detection Based on Speech-Aware Token-by-Token Sequence Labeling with BLSTM-CRFs and Attention Mechanisms.
Tomohiro TanakaRyo MasumuraTakafumi MoriyaTakanobu ObaYushi AonoPublished in: APSIPA (2019)
Keyphrases
- sequence labeling
- conditional random fields
- dependency parsing
- hedge detection
- named entity recognition
- structured prediction
- crf model
- graphical models
- information extraction
- probabilistic model
- random fields
- pairwise
- spontaneous speech
- natural language processing
- higher order
- parameter estimation
- training data
- similarity measure