Active Learning Using Phone-Error Distribution for Speech Modeling
スポンサーリンク
概要
- 論文の詳細を見る
We propose an active learning framework for speech recognition that reduces the amount of data required for acoustic modeling. This framework consists of two steps. We first obtain a phone-error distribution using an acoustic model estimated from transcribed speech data. Then, from a text corpus we select a sentence whose phone-occurrence distribution is close to the phone-error distribution and collect its speech data. We repeat this process to increase the amount of transcribed speech data. We applied this framework to speaker adaptation and acoustic model training. Our evaluation results showed that it significantly reduced the amount of transcribed data while maintaining the same level of accuracy.
著者
-
SHINODA Koichi
Tokyo Institute of Technology
-
Furui Sadaoki
Tokyo Inst. Of Technol. Tokyo Jpn
-
Furui Sadaoki
Tokyo Institute Of Technology
-
MURAKAMI Hiroko
Tokyo Institute of Technology
関連論文
- Acoustic Model Adaptation for Speech Recognition
- Tree-Structured Clustering Methods for Piecewise Linear-Transformation-Based Noise Adaptation(Speech and Hearing)
- Acoustic Model Adaptation for Speech Recognition
- Recent Progress in Corpus-Based Spontaneous Speech Recognition(Feature Extraction and Acoustic Medelings, Corpus-Based Speech Technologies)
- THE USE OF FINITE-STATE TRANSDUCERS FOR MODELING PHONOLOGICAL AND MORPHOLOGICAL CONSTRAINTS IN AUTOMATIC SPEECH RECOGNITION
- Adaptation to Pronunciation Variations in Indonesian Spoken Query-Based Information Retrieval
- Committee-Based Active Learning for Speech Recognition
- Robust Gait-Based Person Identification against Walking Speed Variations
- Selected Topics from LVCSR Research for Asian Languages at Tokyo Tech
- Active Learning Using Phone-Error Distribution for Speech Modeling
- Distance-based Factor Graph Linearization and Sampled Max-sum Algorithm for Efficient 3D Potential Decoding of Macromolecules
- Spectral Subtraction Based on Non-extensive Statistics for Speech Recognition
- Active Learning Using Phone-Error Distribution for Speech Modeling