Improving RNN-T ASR Accuracy Using Untranscribed Context Audio.
Andreas SchwarzIlya SklyarSimon WieslerPublished in: CoRR (2020)
Keyphrases
- high accuracy
- recurrent neural networks
- multimedia
- context aware
- nearest neighbor
- contextual information
- signal processing
- computational efficiency
- highly accurate
- correlation coefficient
- genetic algorithm
- context sensitive
- audio video
- broadcast news
- emotion recognition
- audio visual
- visual information
- speech recognition
- prediction accuracy
- classification accuracy
- decision trees