Multimodal N-best List Rescoring with Weakly Supervised Pre-training in Hybrid Speech Recognition.
Yuanfeng SongXiaoling HuangXuefang ZhaoDi JiangRaymond Chi-Wing WongPublished in: ICDM (2021)
Keyphrases
- speech recognition
- weakly supervised
- object detectors
- hidden markov models
- object class
- topic models
- relation extraction
- language model
- automatic speech recognition
- superpixels
- pattern recognition
- speech signal
- named entities
- supervised learning
- speech recognition systems
- object detection
- semi supervised
- training set
- information extraction
- object recognition
- machine learning
- low level
- pairwise
- multiscale
- computer vision
- multi stream
- information retrieval