LDA-based context dependent recurrent neural network language model using document-based topic distribution of words.
Md. Akmal HaidarMikko KurimoPublished in: ICASSP (2017)
Keyphrases
- language model
- context dependent
- recurrent neural networks
- document level
- document representation
- n gram
- relevance model
- document retrieval
- query topic
- language modeling
- query terms
- ad hoc information retrieval
- information retrieval
- vector space model
- document length
- document ranking
- retrieval model
- probabilistic model
- multiword
- query expansion
- feed forward
- test collection
- neural network
- low level
- word clouds
- text documents
- opinion words
- pseudo relevance feedback
- query specific
- word level
- topic models
- natural language
- context sensitive
- mixture model
- document set
- sentence level
- language modeling framework
- translation model
- expert finding
- smoothing methods
- document similarity
- statistical language modeling
- artificial neural networks
- dependency structure
- document collections
- bayesian networks
- keywords
- news articles
- statistical machine translation
- high level
- retrieved documents
- search engine
- cross lingual
- retrieval effectiveness
- relevant documents