TALCS: An open-source Mandarin-English code-switching corpus and a speech recognition baseline.
Chengfei LiShuhao DengYaoping WangGuangjing WangYaguang GongChangbin ChenJinfeng BaiPublished in: INTERSPEECH (2022)
Keyphrases
- speech recognition
- open source
- source code
- speech recognition technology
- statistical machine translation
- language model
- word error rate
- speaker independent
- automatic speech recognition
- hidden markov models
- parallel corpus
- speech signal
- speech processing
- pattern recognition
- broadcast news
- speech recognizer
- linguistic features
- speech synthesis
- speech retrieval
- machine translation
- noisy environments
- speaker identification
- natural language
- spontaneous speech
- isolated word
- english language
- conversational speech
- parallel corpora
- cross language information retrieval
- handwriting recognition
- cross lingual
- speaker dependent
- information retrieval
- language identification
- speech recognition systems
- language learning
- probabilistic model
- computer vision