Weakly-supervised word-level pronunciation error detection in non-native English speech.
Daniel KorzekwaJaime Lorenzo-TruebaThomas DrugmanShira CalamaroBozena KostekPublished in: CoRR (2021)
Keyphrases
- word level
- weakly supervised
- error detection
- english text
- speech recognition
- error correction
- language independent
- machine translation
- document images
- document analysis
- automatic speech recognition
- word recognition
- relation extraction
- topic models
- n gram
- character recognition
- semi supervised
- sentence level
- hidden markov models
- speech signal
- named entities
- word segmentation
- semantic roles
- cross lingual
- semantic relations
- domain specific
- text mining
- cross language
- noisy environments
- object detection