Improving Noise Robustness for Spoken Content Retrieval Using Semi-Supervised ASR and N-Best Transcripts for BERT-Based Ranking Models.
Yasufumi MoriyaGareth J. F. JonesPublished in: SLT (2022)
Keyphrases
- semi supervised
- automatic speech recognition
- noisy environments
- ranking models
- information retrieval
- learning to rank
- ranking functions
- query dependent
- supervised learning
- semi supervised learning
- document retrieval
- information retrieval systems
- speech recognition
- query expansion
- spoken documents
- image retrieval
- spontaneous speech
- ranking algorithm
- labeled data
- web search
- retrieval systems
- speech signal
- retrieval model
- loss function
- relevance feedback
- active learning
- test collection
- broadcast news
- spoken document retrieval
- hidden markov models
- directly optimize