BROS: A Layout-Aware Pre-trained Language Model for Understanding Documents.
Teakgyu HongDonghyun KimMingi JiWonseok HwangDaehyun NamSungrae ParkPublished in: CoRR (2021)
Keyphrases
- language model
- pre trained
- document retrieval
- ad hoc information retrieval
- query terms
- information retrieval
- document ranking
- vector space model
- language modeling
- n gram
- relevance model
- word clouds
- test collection
- document length
- pseudo feedback
- retrieval model
- query expansion
- relevant documents
- speech recognition
- probabilistic model
- query specific
- pseudo relevance feedback
- document collections
- retrieved documents
- web documents
- training data
- training examples
- retrieval effectiveness
- information retrieval systems
- document similarity
- smoothing methods
- user queries
- language modeling framework
- cross lingual
- training samples
- term frequency
- text documents
- keywords