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: CoRR (2023)
Keyphrases
- semi supervised
- automatic speech recognition
- ranking models
- noisy environments
- information retrieval
- learning to rank
- document retrieval
- ranking functions
- loss function
- speech recognition
- web search
- semi supervised learning
- query dependent
- ranking algorithm
- speech signal
- spoken documents
- retrieval model
- labeled data
- information retrieval systems
- relevance feedback
- active learning
- retrieval systems
- hidden markov models
- broadcast news
- pairwise
- directly optimize
- test collection
- web search engines
- supervised learning
- average precision
- retrieval quality
- search engine
- machine learning